The AI Native Product Manager - Masterclass | HelloPM - Day 1

Indonesia, Himachal Pradesh, Tamil Nadu, Toronto Kanpur Nigeria Ambala UAE Karnataka Haridwar, Indonesia, Himachal Pradesh, Tamil Nadu, Toronto

Yes, we will just start at the right time. So, you guys have to wait for a

time. So, you guys have to wait for a couple of minutes before we can start.

Yes.

Everyone, I can see there are 120 people in the chat right now. Everyone, please

tell me Please tell me which city {slash} country are you logging in from? Yes,

Prajwal, we are going to answer all of these questions today. And tomorrow.

Yes. Everyone, please try to also open your cameras.

When you open your cameras, two things happen.

One is you'll be able to focus more in the session because I believe that you are at one arm's distance from your phone. And if you're not opening your

phone. And if you're not opening your camera, there is no reason for you to go ahead and focus in the session, right?

The second part is Yes. And the second part is you need to

Yes. And the second part is you need to if you go ahead and Yes, if you go ahead and open your cameras, I'll be able to feel more interactive in the session.

Just give me a moment.

Yes. I think we are right on time. I

hope everyone is able to see my screen, right?

Yes, on the screen, let me write something which is the topic for today, which is AI native PM.

Right?

Yes.

Cool. So,

now everyone Everyone, welcome back.

Yes, I hope everyone is able to see AI native PM on the screen, right?

A quick yes or no in the chat before we can go ahead. Yes, cool. So, welcome

everyone. This is the session on understanding Yes, I'll stop the chat for some time so that you guys can focus. I'll keep on opening it in midway when I ask for your opinions.

So now, everyone listen to this carefully.

Today's session is going to be a bit different from what you might have seen in the free master classes from the HelloPM in the past, right? In the past master classes, what

right? In the past master classes, what we used to do is I would have a slide deck.

I will give you a walk through of the slide deck with the help of multiple case studies and examples.

But now, what we have done is we have put in almost all of that content on YouTube.

Right? More details of that content are available in the cohort, but all the free part is already available on YouTube.

So, I thought that today we are going to take a different approach.

Today's whole session is going to happen on the whiteboard.

I want your support in order to discuss the items. So, what we'll not do is that I'm only speaking and you are only going and listening to it. We are going to keep the session interactive. Right? We

are going to go ahead and build up this content along with each other so that you are able to get the most advantage from the session.

Right? So, I want your support. Now, as

per the agenda for today, what we are going to do is Yes, just one moment.

Yes.

What we are going to do is we will just go ahead.

We will just go ahead and first we will talk about what or who is AI

native PM.

Right? After that, we will talk about some use cases of how different kind of PMs are using AI in their workflows.

And after that after that, I'm going to tell you how you can use top tools. I'm actually going to give

top tools. I'm actually going to give you a walk through. Do you understand?

The tutorial videos are already available on YouTube.

Right? I'm here to show you some kind of cases.

Right? So, this is going to be the agenda for today.

Tomorrow, we are going to go a bit more advanced. And tomorrow, we are going to

advanced. And tomorrow, we are going to talk about how the large language models work, how do you build products on top of that, how a truly AI native PM works.

Right? So, today is a fundamental session. Today is Tomorrow is a more

session. Today is Tomorrow is a more advanced session. The recording of this

advanced session. The recording of this session will be shared with everyone along with all the resources who are remaining till the end and sharing the feedback form in the end.

Right? So, this is what we are going to go ahead and do today. Okay, before I can start before I can start, I want one help from everyone. Everyone,

everyone. Everyone, please go ahead and write in the chat.

Please go ahead and write in the chat.

What do you think AI native PM is? What

is your definition of AI native PM?

Send in the common chat, please, so that everyone can read. Don't send it only to me.

Okay, please do not share the links on the WhatsApp group. Not recommended.

Yes.

Cool.

Very interesting definitions.

Right?

Very, very interesting definitions. Let

me see what we can do together.

Okay.

So, someone told Yes, first is someone who understands AI and make AI updated decision based on customer needs. Okay. This is given by

needs. Okay. This is given by Anshul.

Second is is a very simple from Vani.

Not a simple one, but a very interesting one.

Which is those PMs who can use AI in their thinking as a thinking partner and can also visualize how product can be evolved using agentic AI warning working behind the screens.

Yes, Gurvinder has a more elaborate definition.

Which is someone who can actually build a product with AI, not by just using AI features.

Right? And then according to this, we also have hundreds of other people who have gone ahead and given their inputs, right?

So right now we have 500 people in the session. We are maxed out on the number

session. We are maxed out on the number of people.

Now, I have got very interesting insights, right?

Now guys, before I could tell you the real definition, please look at this very carefully.

Okay, I'm trying to help you understand what a product manager does regardless of the company, regardless of AI, regardless of technology, regardless of the domain. Okay, listen to this carefully because this is the actual

foundation. No matter AI is there or

foundation. No matter AI is there or not, this will always remain true.

Right? So what happens is as a product manager, your company comes to you and tells you that this is our business goal.

If you are already a senior PM or a director PM or a VP PM, then you might be responsible for creating this business goals as well.

But for most of the use cases, the company will give you the business goal.

Right?

Now based on that business goal, you are going to do some kind of brainstorming.

You are going to understand what is the kind of market that we are operating in.

You will understand what is the kind of market that we are operating on.

You will go ahead and collect a lot of data.

You will go ahead and talk to your stakeholders such as sales and marketing and other people. Why? Because they are also closer to the customers.

Stakeholders.

Right? In terms of market, you are going to look at trends.

And you are also going to look at your competition.

Right? So business goals, market understanding shareholder understanding, you are going to have all of this.

Plus as a PM, you are also going to do some kind of brainstorming based on whatever information is available to you from your own mind.

Right? Based on your experience, based on your own work.

And then you'll also collect the data.

Based on that, you will have certain ideas.

Or I would say certain kind of opportunities.

That we can do this, we can do this, we can do this and there could be a very long list of things that you can do as an PM.

Right? Now, all of these opportunities are not good enough to pursue.

A PM always has an issue of plenty called as a problem of plenty, which is you have more stuff than you can actually do.

Right? You will have always more backlog items than the number of engineers or designers that you have in the team.

Right? So in order to make sure that you are only capturing the right opportunities, what you will do is you will just go ahead and understand and validate these opportunities.

How do we validate the opportunities?

Can you guys please tell me in the chat how do we validate the opportunities?

Let's say you have a list of opportunities.

How do you know which opportunity to pursue? Tell me in the chat.

pursue? Tell me in the chat.

Everyone, we are trying to keep this session very interactive, so please be interactive in the chat.

Yes, very good.

Yes guys, so what we are going to do is we are going to understand these opportunities from all the lenses.

We are going to understand if the market wants this, we'll do the user research and market research. We will understand it from the business perspective whether this will help the business or not. We

will go ahead and look at the data. We

will talk to the stakeholders. We will

implement multiple principles and multiple research methodologies in order to validate this.

Right? So in the validation, you are going to again do all of these things in order to validate the idea.

Right?

And after validation, what we'll do is understand we are validating the idea by doing user research and everything because code is costly.

Right? If you have do not have a validated idea, if you give it to the engineering team, if they develop it and they are able to find that no, it is not needed, then you have wasted the engineering efforts and some time for the company.

Right? Now after validation, after the ideas are validated, you are going to prioritize them.

Right? You can do a simple prioritization based on what is going to be the impact of this feature and what is going to be

the cost of building this feature. Cost

as in resources. How many engineers are required, how many time is required, how much money is required. Impact is the user in the business impact.

Right? After this, once you prioritize this, you will have something called as a product roadmap.

Product roadmap is nothing.

Product roadmap is nothing, but is a set of prioritized ideas or initiatives that are going to shape up your product.

Right? So Claude code, like Claude initially released chat.

Then they released code.

Then they understood that Claude code is only for technical people, so they also released code work.

And now in order to compete with lovable and Google slides of the world, they have recently launched Claude design.

Right? So

a document where it is written that first Claude chat, then Claude code, then Claude code, then Claude code work and then Claude design, that document would be called as a product roadmap.

Right? If you guys want to understand what roadmaps are, how they look like, a better way is you can just go to let's say this com- let's say Claude

release notes.

Or you can take any company and write release notes, you should be able to find how they are releasing.

Right? You should be able to understand their roadmaps, how they are evolving.

Right? So

the first use case of AI that I'm going to tell you right now for product roadmap is very simple. Right? What you

can do is if you want to understand our competitor, take their release notes and go to Claude and ask, can you convert these release notes into a good

well-structured roadmap?

Let me show an example.

So if you want to understand your competitor's roadmap without going on to one-on-one release notes by yourself, you can do this.

Look at this. I'm not sure if you guys are able to see this, but look at this image, you should be able to find it on my LinkedIn profile. So what I've done is I have gone ahead and given a simple

prompt to Claude.

That this is the URL.

This is the URL for Claude release notes.

Go ahead and give me the last 60 days roadmap.

Right? And then I was able to understand, so if I'm a competitor to Claude, I should be able to understand where their strategy is actually leading.

Right? This thing was almost impossible few years ago because you cannot go through so many documents at your will.

Right? And you can only do it for one or two competitors. Now you can do it for

two competitors. Now you can do it for any number of competitors.

Right? So this is the first part. Now,

Yes. Now that we have Yes. So we will have a product roadmap.

Yes. So we will have a product roadmap.

Once we have the product roadmap, understand product roadmap is high level. Like by looking at this document,

level. Like by looking at this document, you cannot give this document to your team and tell that you should go ahead and develop this, right? You have to give a more detailed document to your team,

which is called as product requirement document.

Right? So you are going to create something called as a product requirement document, which is nothing but a piece of paper or a Google Doc or a Confluence Doc where you are going to

write what this feature is all about.

Why are we building this?

And this will give the alignment to the team. Right? For example, if you have

team. Right? For example, if you have four products in the roadmap, for each of the product, you will create a product requirement document. If the product is too complex,

document. If the product is too complex, for example, Facebook, then every feature can have its own PRD, product requirement document.

Right? If you guys do not know how to create a PRD, I will share a video resource with you so that you can understand within 10 minutes how to create a product requirement document.

That is not rocket science. That has

become much simpler with AI.

Right? So, this is the product requirement document.

Now, after the PRD after the PRD, what you will do is you will go ahead and create something called as user stories.

You will create tasks.

And you will create something called as a backlog.

Right? So, product requirement document is about a particular feature or a product. But now user stories are

product. But now user stories are simple, unique tasks that I need to do these 10 things in order to build this product. And this is shared with

product. And this is shared with engineers so that they can pick a task, they can do it, and then they can update you.

Right? This work so far has been done by a product manager or a product owner.

Right? After this, and understand guys, user stories are written in a particular fashion. There's a template for user

fashion. There's a template for user stories that as a user, I want this so that I can do this.

Right? And this needs to follow a particular format to be helpful to the engineers. Right? Once this is done

engineers. Right? Once this is done the engineers are going to execute. Execute means they are going to

execute. Execute means they are going to do the coding, they are going to do everything, and after the execution is done the product is going to

launch. Launch means

launch. Launch means you are going to take the help of GTM, a go-to-market strategy. GTM means

go-to-market strategy. GTM means go-to-market strategy, where you are going to take the help from the marketing team plus the sales team.

And they are going to help you ship this product to the market. By market, I mean they are going to help you ship this product to the users of the product.

Right? And once the users have seen the product the users are going to give you the feedback so that you are going to improve the product again.

And users are also going to use the product so that you can collect the data, and then you can improve the product.

Right? And this cycle will keep on going on.

And this cycle will keep on going on, and then you will go ahead and understand what is a product what is a product development life cycle.

Right? So, in this cycle, we have three phases.

The first phase the first phase or this phase where we are creating a product roadmap document this is called as

the discovery phase.

The next phase where you are building the product is called as the delivery phase. And the

third phase where you are actually going ahead and shipping the product to the users, launching, this is called as

the distribution phase.

Right? So, this is what a product manager does regardless of the company.

Understand in some companies you'll be focusing more on discovery, in some companies you'll be focusing more on distribution, in some companies you'll be focusing more on delivery. Okay? But this is overall the

delivery. Okay? But this is overall the encompassing role for a whole product manager's job.

Right? Everyone, quickly in the chat tell me yes or no if you're able to understand this view.

Yes.

Yes. Now

let's talk about the AI part. Okay? So,

understand guys before you become an AI PM, you have to become a PM.

Right? Before you become an AI PM, you have to become a PM. So, you cannot skip the basics. The number one mistake that

the basics. The number one mistake that people are doing these days is they are straight away jumping to the tools.

Right? But if your fundamentals are not correct, you cannot give an exam of class 12th before you passed grade one or grade two. Right? So, it is important to understand the fundamentals first before so that I can go ahead and tell

you that.

Right? Now

this is mostly the pre-AI phase.

Right? This is how product management used to happen. In the discovery, you have to go ahead and put a lot of efforts in order to talk to users and everything.

And then once you are confident enough once you have a lot of data then you are going to go ahead then you are going to go ahead and give it to the

engineers. Why? Because engineers'

engineers. Why? Because engineers'

bandwidth is very costly.

Right?

Now everyone, people who are PMs in the room tell me Okay, this question is for everyone.

Tell me everyone in the discovery, distribution, and the delivery phase ideally as a PM where you should be spending most of your time.

I will repeat the question.

As a product manager in all of these three phases, discovery, distribution and delivery, where as a PM you should be ideally dedicating most of your time.

Yes. So, discovery, right? Now

pause everyone.

Pause everyone, okay?

Tell me like people who are PMs, they should only answer this.

Tell me by putting a hand on your heart very honestly where do you go ahead and spend your time mostly right now?

Right? So, now you guys are able to understand where we are taking this.

Right? You want to spend time in discovery, you know that's that's it you know what is good for your job. But

you're not able to do this. Why? Because

you are spending your evenings in order to look at what engineer has done.

You are spending your mornings in order to connect with these stand-ups. And you

are spending all of your days in order to just go ahead and connect with the engineers and the designers rather than talking to the customers.

Right? So, that is a sad reality of product management right now.

Right? And this is where AI plays a very important role. Now, what is happening

important role. Now, what is happening is because we understand because we understand that delivery is difficult and we have to deal with some difficult people during the delivery

and there are always going to be expectations from people that we should be able to ship this feature as of yesterday.

The problem is that because of too much pressure on delivery because of too much pressure on delivery PMs are taking very mediocre and very

safe decisions.

Right? PMs are taking very mediocre and very safe decisions. Now, why it is happening is because understand I told you that there is a framework called as cost and impact.

Right?

Do you really believe that if for a for a really ambitious idea you'll be able to actually understand what is going to be the impact and what is going to be the cost?

Right? It is actually very difficult, but understand it is not your company.

It is not your company. You are

responsible for some kind of numbers.

And that is why you are not able to take some very risky decisions because you don't have data. I'll not call it risky, I'll call them high-leverage decisions.

Right? Because of which every company is just trying to compete with their, let's say, copy their competitors. Every

company is trying to play mediocre by taking the safest decisions.

Right?

Because these safe decisions are taken because all of this cycle it actually take months right now.

It actually takes months. Okay? So, if I go ahead and give you a glimpse Yes. So, if I go ahead and give you a

Yes. So, if I go ahead and give you a glimpse in discovery what you guys need to do is Yes.

In discovery, what you guys need to do is actually a lot of things. Okay?

You have to talk to users, before that you have to find the users, you need to

create user profiles, you need to find problems, you also need to create prioritization.

Right? And you're also responsible for creating documentation and multiple alignment

meetings and a lot of market research.

Right? Similarly

similarly in the discovery part in the delivery part alignment with engineers documentation meetings

more meetings more documentation Right? And

Right? And then follow-up meetings Right? Because someone else is doing the

Right? Because someone else is doing the job, there is some kind of knowledge transfer that is going to happen.

Expectation setting has to happen. What

happens is you end up doing many meetings that you should actually do.

Right? And most of these meetings are just to help the other person understand what you're trying to do, and then managing the timelines.

Right? And apart from this in the distribution part in the distribution part, what you will do is you will take the help of the product marketing and the sales team in

order to create a GTM strategy finding your ideal customer profile finding right communication

designing landing pages release notes analytics and more meetings

Right?

Everyone tell me agree in the chat if you agree with these tasks that happen in the product management world right now. If

you as a PM has done these jobs, tell me agree in the chat.

Yes.

Yes now Now understand.

As a PM as a PM there is so much potential for you to reduce this crap reduce so much of this crap by

effectively using AI which is the purpose of an AI PM, AI native PM's job. Okay. So, AI native PM is not someone who is just going ahead and using some AI tools.

AI native PM is not someone who is just building with AI.

AI PM PM AI native PM is not someone who is just going ahead and utilizing AI to build a product. Okay.

AI native person is actually a profile where someone is able to understand the roadblocks in their entire PM cycle and then they are able to use the right

kind of tools in order to not only become efficient not only become efficient, but in order to get better impact.

A common mistake that people do is they use AI in order to save time.

Right? They use AI in order to save time. For example if you are taking 6

time. For example if you are taking 6 months in order to create a product that is going to give you $1 million in return and if you are trying if you're looking

that I can go ahead and do that job in 1 month, but the product is not giving you $1 million, it is giving you way less, then you have failed as a PM.

Right? So, as a PM you have to understand you're not focusing on the output, which is how much time you have saved, how much product that you have created. You have to focus on the

created. You have to focus on the impact.

AI native PM is a person who understands what are the roadblocks in their journey as a PM in order to create a good product. And then they go ahead and

product. And then they go ahead and leverage AI in order to solve these problems. The problems could be with their PM workflow, could be with their customer

workflow, could be with their product, could be with anything.

Right? Now, going forward let me go ahead and give you few more items. Right? So, this is discovery, delivery, and distribution.

Can you guys help me understand how can AI help with these three things?

Go ahead, guys. Think harder for yourself before I could give you the answer or an approach. How can AI help me in these three things?

Talk to users, find users, create user profiles.

Go ahead, guys. Think harder. Take a

minute. No rush. Take a minute in order to give the answer.

I have all the time.

Yes.

So, there are so many things that AI can do.

Right? I'll show you one example. Okay.

Look at this very carefully.

So I personally use this tool called as Cursor and Claude Code. Okay. Look at

this carefully, guys.

Right? So, where is the folder?

Yes.

So look at this carefully now. Okay?

I'll give you one expression and then you'll understand what I'm doing. Okay?

So, let's say this is Cursor opened on my computer.

Everyone, a quick yes or no if you're able to see Cursor is screen open.

Yes. Thank you.

Yes. So, what I've done is I've installed Cursor and I've installed Claude Code on my computer.

Right?

Now, what Claude Code does is as you can go to Claude Chat, you can ask a question and then it will give you the answer.

You can also attach a couple of files.

But in Claude in Claude Chat when I use on browser, I have to keep on attaching files again and again.

In Claude Code, what I can do is I can just give it a folder that this is a folder and then you go ahead and do your work.

Right? So, it is more transparent and I would say a more flexible tool as compared to Claude Chat. Okay. So, in

Claude Code, what I've done is I have given it few folders.

So Yes.

So, at Hello PM whenever any cohort starts any cohort starts, we do a kickstarting call with everyone.

Kickstarting means people would come and they will introduce themselves.

Right? So, the problem that I'm trying to solve right now is let's say I want to improve the landing page of Hello PM.

Or if I want to improve the course content of Hello PM.

Or if I want to improve any other product aspect of Hello PM.

Right? If I want to improve or change anything in Hello PM if I want to change or take a decision at Hello PM I need to make sure that I'm thinking like a user

and not like a seasoned product manager.

At Hello PM, we have people who are either coming from non-product backgrounds or they have about 5-7 years of experience in product management.

Right? I have more experience. So, it

might be possible that some decision that I'm taking, they are not too obvious or maybe good for my students.

So, I want to make sure that I am very much user-centric in the program.

Right? I hope everyone is able to understand the problem. Tell me yes or no if you're able to understand the problem.

I'll repeat this again.

I want to take the decisions about the program based on the based on my users

profiles.

Right? I want to have more user empathy.

Right? So, now what I've done is in order to solve this problem, what I've done is Yes. What I've done is

Yes. What I've done is rather than randomly saying AI that these are my users because understand, like you, I'm also lazy. Maybe I'm more lazy than most of you in the room out

there. Okay? So, I cannot go ahead and

there. Okay? So, I cannot go ahead and with full focus I can tell AI that these are the four people who are going to join the program. Even if I say this, my information would be incomplete.

So, what I've done is at Hello PM, whenever a person joins in the first call, like when every batch starts, in the first call, we ask people

why have you joined Hello PM program?

What is your background?

What do you want to get from this program?

Right? Based on that, what we do is we do a call. And in this call, about 100-200 people go ahead and they introduce themselves. So, what I've done

introduce themselves. So, what I've done is I have gone ahead and taken these call recordings in the form of transcripts for cohort 47

Look at this.

Cohort number 47, 48, and 49.

Right? This has data for almost 400 people who have introduced themselves.

You can also find these YouTube videos on YouTube videos. Right? We have made this information public. Right? Now,

what I've done is I have gone ahead, asked Claude code that read these three files,

and give me seven user personas which are able to cover all the aspects of these users. Okay? And then it was able to give me these files. Look at

this file persona. So, first is the industry is the service industry SKP.

There are many of you who work in service industries. Now, you want to

service industries. Now, you want to move into a product-based company as a product manager. Right? So, this is a

product manager. Right? So, this is a detailed about of this profile.

Second is senior AIPM. Person who's a senior AI senior PM, but they now want to understand into AI.

Right? Third is a person who is a global audience.

Right? He's an Indian-origin person, but he is going ahead and right now living in the US.

Right? And maybe there could also be aspiring builders, someone who wants to go ahead and build a product out there, founder in making.

Right? So, I have created these detailed profiles with the help of Claude.

And next time if I want to make any changes in the program, I'm going to ask Claude that this is my decision. Can you

vet this decision based on all of these profiles?

Right? This will make sure that I'm not being too short-sighted, or I'm not thinking only from my perspective, but I'm having genuine user empathy.

Right? And because I know that I'll be limited in my thinking, I have also asked Claude that please go ahead,

please go ahead and tell me what are the other ways in which I can go ahead and use these personas.

So, it has also given me a lot of other suggestions where I can utilize these personas.

Right? Understand, guys?

This use case that I have used, it was not written in any YouTube video.

It was not mentioned in any YouTube videos. It was not written in any PM

videos. It was not written in any PM playbook.

Because we faced the problem, we were able to identify the solution.

Right?

Tell me, guys, in the chat, did you find the solution interesting or helpful?

First, tell me if you're able to understand this.

Yes, seven is an arbitrary number, guys.

Three, four, I can also do.

Beyond seven, I would not be able to go ahead and take better decision because it would be diluted. Right? So, seven is an arbitrary number. You can try five to seven, that's a good enough number.

Yes.

Cool. Very important point, Varun.

Right? Guys, my recommendation is I have done it with Claude code, but my recommendation to everyone is please do not fix on a tool. This is

what I'm trying to teach you today.

Okay? You do not have to think that only you should use Claude code, or maybe notebook LM. Right? You will

use these tools in order to figure out what is good for your use case.

Right? So, try notebook LM, try ChatGPT.

After using it for a couple of weeks, you'll be able to understand which is better for your use case.

Right? Please do not stick with a tool.

Right? Now,

this was the use case.

Now, let me tell you how to exactly think about this.

Understand guys?

You have two options.

Actually, three options. One is

you can go ahead and not use AI in your workflow at all.

Right?

You can have you can not use AI in your workflow at all.

In that case, you are basically leaving money and opportunity on the table.

Right? A common thing that people tell me is a common worry that people tell me is that AI hallucinates.

Right? Everyone tells me that we are using so much AI, but I'm very much afraid. I'm very much afraid to

much afraid. I'm very much afraid to learn AI and use AI because AI hallucinates.

Okay? Can anyone answer me this question that if AI hallucinates, what should we do? Should we not use AI?

Tell me, you guys are product managers, right? I am throwing a problem statement

right? I am throwing a problem statement at you.

Right? So, AI hallucinates, all of us knows.

So, how do we go ahead and make sure that we are still able to create products on top of hallucinating AI?

Yes, very good. Guys, understand.

The answer is very simple. Okay?

You should first go ahead and write some kind of evaluations. If you guys don't know about evaluations, search for Ankur Shukla evaluations. I have done a

Shukla evaluations. I have done a podcast with Akash Gupta. You should be able to understand them clearly. Okay? A

technical answer is evaluation.

The second answer is designing your prompts better.

The third answer is AI is evolving. So,

the the amount of hallucination that used to happen couple of years ago, they are right now not even 10% of that.

So AI is improving.

Right? But understand, these are technical answers.

People people would still be able to find loopholes or gaps in this.

Right? What you actually need is not just the answer for hallucination. That's a technical thing.

hallucination. That's a technical thing.

You will ask AI that how do I reduce hallucination in this model, and it will tell you how to reduce hallucination.

The answer is not that how do you reduce hallucination. The answer is actually

hallucination. The answer is actually the mindset shift.

Right? Imagine

when this alternating current was invented, you know that every year we lose so many lives because of alternating current.

People get shocked, they get electrocuted. Many animals, many humans

electrocuted. Many animals, many humans go ahead and die every year because of electrocutions.

Right?

What if we did not have invented AC, like electric like alternating current, or we were too cautious to use electric current at all the places?

The humanity humanity would not be able to go ahead and move ahead.

Okay? So, whenever any kind of technology comes, every technology is like fire.

Right?

It can go ahead and do a lot of things for your advantage. It can take the humanity forward.

And it also has some kind of it also has some kind of bad behaviors.

Right? For example, fire can help you cook your food.

And fire can also wipe away an entire forest.

Right? The purpose is that as humans, we have learned how to control these things.

Right? So, first you need the mindset.

If you are if you are someone who is, let's say, a a bit of pessimistic about AI, it is only your loss.

Why? Because you are not able to leverage an important piece of technology.

Right? And then what will happen is that you guys are going to pay a cost called as an opportunity cost.

Which is rather than learning how to use AI, you just keep on finding the mistakes in AI. And then what will happen? You will

AI. And then what will happen? You will

go ahead and understand that I could have gone ahead, tried to solve this problem, rather than be a part of the problem.

Right? So, everyone, can you tell me in one line what did you learn about in the last 5 minutes?

Everyone, do active recall. Tell me in one line in the chat what did you learn in the last 5 minutes?

Yes. So, you should learn how to use AI.

I'm not saying that you should blindly follow what the AI says.

There are multiple things that are already there. Right? I have created a

already there. Right? I have created a detailed masterclass which many of you might have watched on YouTube, which is a 3.5-hour masterclass, where I have talked about how to control hallucination, some part about

evaluations, prompt engineering, rag.

All of these are proven methods to reduce hallucination. They do not take

reduce hallucination. They do not take it to zero.

Right? Understand every technology that you're sitting in these days. For

example, the laptop that you're working on, the internet that you're working on, they also have some missing links.

Right? You can lose your data from your hard drive.

You can go ahead and break the internet connection while a call is going on. But

that does not mean that people are are going to use the technology. It is only your loss if you do not understand how to work with AI. Because businesses are already Businesses have already taken a call.

Right? So, businesses have already taken a call. It is your call either you want

a call. It is your call either you want to leverage the AI and move forward with that or you keep on crying that no AI is hallucinating. I cannot go ahead and use

hallucinating. I cannot go ahead and use it for all the purposes.

Yes.

Very good. I hope everyone is able to understand this mindset. Right? Now,

moving forward.

Okay. Now, moving forward, we have talk about So, now everyone look at this carefully. I'm going to give you a very powerful framework today. Okay? So,

today. Okay? So,

I was talking about three parts. The

first part is There are three kinds of people in the world right now.

The first kind of people are AI is just hype.

Right? The first kind of people are AI is just hype.

I understand them. I understand where they are coming from, but it is very harmful to be in this part.

Right?

The second kind of people are people who are relying on someone else to solve their problems. Right? So, people

Right? So, people like the people in this room who are relying on me to tell them that how should you go ahead and use AI.

All of you are a part of the second part, which is need slash looking for pre-created solutions.

For example, you need a prompt to create You want to use AI to create a product requirement document, but you need a prompt.

You need a use case. You need to You need You need AI in order to do customer research, but you want to see my workflow.

Right? These are the people who are learners by observant by being observant. Right? So, you want a

observant. Right? So, you want a playbook where I can go ahead and teach you that this is how you are going to use AI.

Right? And you guys are going to YouTube. You are going to chat GPT in

YouTube. You are going to chat GPT in order to figure this out.

Right? But, understand, guys, this is These people are, let's say, the bottom 90%

These people are the top 10%.

Right? But, the two people the third grade that I want to give you in, which is the top 1%, are the people who are able to

figure out how to use AI in in their situation.

You are able to invent, discover how can I solve my problems at scale with AI.

Right? So, the use The workflow that I've told you that I was able to collect my data from my students. I was able to give it to AI. Then, I was able to create personas. And then, I was able to

create personas. And then, I was able to create decisions. That is something very

create decisions. That is something very unique to me.

Unless you are able to build this kind of muscle, I can assure you you'll just be in maybe the top 10%, but not the 1%.

Right?

Everyone tell me yes or no in the chat if you're able to understand this aspect. The three buckets if you're able

aspect. The three buckets if you're able to understand.

Yes. So, guys, everyone is saying yes.

Now, tell me how do you get into from second to third? I'm sure everyone is on to two.

third? I'm sure everyone is on to two.

If you're joining this session, you are at least not a one.

You are a two.

At least a two. Right? Now, can you guys tell me how do you move from two to three?

No, guys. You don't need to join Hello PM for that.

Yeah. At Hello PM, we definitely we are going to give you 10x deeper content. We

are going to give you hands-on understanding of everything.

But understand if you have the right kind of mindset, you should be able to do it by yourself as well.

Okay.

Yes. So,

guys, I'm going to go ahead. So,

understand, this is possible.

You just need a framework.

Okay.

Framework is the answer here. Although,

I'm not a personal lover of frameworks, but here we have created such an interesting framework that is definitely going to help you take your AI skills to a different level altogether.

Okay. Let me tell you this.

So, the name of the framework is Maybe I'll write it down because my handwriting is not so good.

Yes.

So, the name of the framework is P O W E and R.

Also called as power framework.

Right? Now, what

Now what power stands for?

The P in power stands for The P in power stands for possibilities.

Possibilities.

Right?

The O in power stands for opportunities.

The W in power stands for workflow.

And the E in power stands for engineering.

And R in power stands for reflection.

Right? Possibility, opportunities,

workflow, engineering, and reflection.

Right? Now, what we'll do is before I could help you understand this framework. Understand, guys,

framework. Understand, guys, trust me when I'm saying this.

You only need this one framework in order to become an AI native professional.

Whether it's a PM, marketer, sales engineer, anyone. You just need to

engineer, anyone. You just need to understand this framework.

Okay. So, I'll leave you to this cliffhanger right now. I'll take a break of 5 minutes. Everyone, get up from your chair. Drink some water.

chair. Drink some water.

Come back in 5 minutes, and we are going to restart the session resume the session and discuss more about this framework when this timer ends.

Go ahead, guys. Take a break.

Hello everyone.

Welcome back.

Now guys, I have a simple request with each one of you.

That if you're enjoying the session, if you're liking the value in the session, I would really appreciate if you can share your quick two three word feedback in the link that I've shared in the chat.

We are trying to keep this free for everyone and your support will really help us. So, please go ahead

help us. So, please go ahead and share your feedback in the chat. If

you share the feedback, what will happen is that other people will get informed about this content and they can also go ahead and consume this content for free, which will help us help more people.

So, go ahead. I have written I have posted a link in the chat for LinkedIn.

Go ahead and write your comment over there. How you're feeling in the

there. How you're feeling in the session. If you're finding the session

session. If you're finding the session valuable, please let me know in the comment of that LinkedIn post.

Give reaction to the message once you're done.

Go everyone. Take a couple of minutes.

Take 1 minute in order to do this before we start.

Everyone, I'll give you maybe 20 more seconds to do this.

Go ahead. Take 20 seconds.

Give your comment on the LinkedIn link that I have provided. It would really help us help more people.

Wonderful.

Come back everyone.

Now, please put all of your focus in the session. This next 10 minutes are the

session. This next 10 minutes are the most important part of the session.

Right? Next 10 minutes are the most important part of the session. Right?

Look at this.

So, I was talking about the power framework.

Right? I was talking about the power framework.

The power framework actually contains five major pillars or steps. Okay?

If you want to really become an AI native professional, these are the five steps that you need to follow.

Right? The first step is possibilities.

Possible Possibility means you should understand what AI can actually do.

Right? Before going ahead and finding what you can do with AI, you have to understand what AI is actually powerful enough to go ahead and do.

Right? So, can you guys tell me what AI can do?

What can AI do? Tell me in the chat.

Guys, let's be productive in the chat.

Tell me what AI can do.

Yes.

Correct. So, understand,

if I have to give you in simple structured part, AI can actually do three things very nicely.

Okay? Different models have different kind of uh uh uh scale of doing this nicely, but there are three things that AI right now is able to

The first part is the AI is able to understand.

The AI is able to understand. So, if you give it any data, the data could be text, image, video, score, it should be able to understand this data.

Right? Eventually,

any video, image, audio, it is actually some data.

Right? So, AI is able to understand the data that what this data is all about like a human.

The second part is AI can go ahead and it can transform the data.

So, if you give it a whole page or maybe a whole book to AI, it should be able to go ahead and summarize this.

It should be able to contain into a blog or some other format.

So, it should be able to transform the data.

And the third part is it should be able to generate the data.

So, if you give AI a simple prompt that write an essay on AI native PM, it will go ahead and generate that for you.

If you ask it to write a code to do something, it will go ahead and generate it for you. So, understand, transform, and generate.

Right? Now, this is actually like going ahead and having a brain. So, AI can go ahead and solve

brain. So, AI can go ahead and solve problems. It can generate content. It

can understand content. It can also go ahead and transform content.

Right? Now, based on this, you can understand what are the different possibilities with AI. Some people are summarizing their meetings.

Some people are getting insights from the meetings. Some people are actually

the meetings. Some people are actually using AI in order to not only summarize or getting insights, they are actually creating to-do items from their meetings. So, I am doing a meeting with

meetings. So, I am doing a meeting with you guys. I give you a homework today.

you guys. I give you a homework today.

Then automatically, the AI will understand this homework and it will give it to you on the day of deadline.

It will mark it in your calendar.

So, what AI can do is limited by only your imagination.

Right? So, this is for AI. Right? I'll tell you about

for AI. Right? I'll tell you about imagination. Okay? So, I'll I'll give

imagination. Okay? So, I'll I'll give you a a good example.

Where you have to think out of the box.

Okay? So,

I'm sure many of you might have used AI in order to generate images or infographics.

Right? In images and infographics, AI generally makes spelling mistakes.

Right? A quick yes or no everyone if you're able if you are if you agree with this. AI generally makes spelling

this. AI generally makes spelling mistakes in images and infographics.

Right? And the infographics look very AI generated.

Right? So, now

you can either wait for the model to improve or you can keep on cursing the AI that no AI is not able to do my job. I have

to do the do the job manually. Or you

can go ahead and think out of the box.

Okay? Which is AI cannot generate images and infographics really well. Not at

least not every AI. Okay? But every AI is able to generate HTML and code really well.

Right? AI is able to go ahead and code at scale.

So, what I do is I ask AI not to create an image or infographic. I ask AI that why don't you take this data and create a detailed infographic for me

in the HTML format.

Right? And then I go ahead and take the screenshot of that HTML. And now I have the infographic which is without any spelling mistake. And if I want to

spelling mistake. And if I want to change anything, I can change it in a couple of minutes.

Right? A quick yes on everyone if you're able to understand this.

Right? So, what you need to do is understand AI. AI is able to create text

understand AI. AI is able to create text content really well.

Claude, Gemini, everyone. Although

Gemini says or ChatGPT says that they'll be able to create good images, but there are some of the touch of AI.

Plus, it is doing some spelling mistakes here and there. So, rather than waiting for AI rather than waiting for AI to do all of this, the better approach is go ahead

create a HTML page create HTML page and then take a screenshot. So, the example that I was

screenshot. So, the example that I was showing on LinkedIn this was not actually an image.

This was an HTML and I have taken the screenshot.

Right? So, what I'm trying to tell you is guys understand you would have seen that there are similar kind of tools

that are available to almost all the people in the world.

For example you have availability of Photoshop. You can

spend a couple of thousand rupees or maybe uh uh $20 in order to get Photoshop. And there is also a

Photoshop. And there is also a professional person who is using Photoshop.

Right?

Even if you are using the same tools the kind of creativity and the excellence that you both have is entirely different. He's a professional.

entirely different. He's a professional.

He can do so many things because he's he is into that tool out there.

If you really want to leverage AI, please do not depend on someone to tell you how to use it. You have to apply your own mind. This hack

no one gave me.

I was able to figure out because I had that problem that I needed to solve.

Right? So, what you need is curiosity ability to tweak. That is how you are going to learn.

Okay? Forget about getting a playbook where everything is written for you.

Even if that is written, if it is available for everyone, then it will cease to become your advantage.

Right? So, this is the first part which is the possibility.

You first have to understand what AI can do.

Right? You have to understand what AI can do.

The second part is also uh uh uh One idea to understand this is Okay, I'll repeat the framework once again, guys. Let me go ahead and

again, guys. Let me go ahead and complete that. Okay? So, first idea is

complete that. Okay? So, first idea is the possibility. You have to understand

the possibility. You have to understand what AI can do and you can figure this out by first using multiple tools and writing on a piece of paper very

carefully that what AI is been able to Right? First part, possibility.

Right? First part, possibility.

Everyone, yes or no if you're able to understand what possibility means.

Yes.

My case was very simple, guys. I wanted

to create infographics with AI. But AI

was making mistake, plus it was taking a lot of time. And whenever infographic was generated, I cannot edit it easily.

So, what I've done is I asked Claude and GPT in order to create an HTML page.

Right? In a particular design. And then

once that HTML page is generated, because it is text AI does not make some mistake. Right? And if I want to change

mistake. Right? And if I want to change it, I can just right click right click and change it or I can ask AI to change this.

The change is much cheaper than editing a infographic.

Right? And then I can go ahead and take the screenshot.

Right? So, first possibility what AI can do.

Second is opportunity.

Now, this is an important part.

Right? This is an important part.

Opportunity means opportunity means where is your opportunity?

Which is what is happening actually in your company where you are spending a lot of time. Where do you think AI can be used? For example, everyone told that

be used? For example, everyone told that we are we are required to spend a lot of time in delivery. So, you can take the

in delivery. So, you can take the delivery. You can understand what are

delivery. You can understand what are the actual things where you can go ahead and use AI. So, understand an opportunity can be used in two ways.

First is can be assessed in two ways.

First is what is the time and what is the accuracy that you need?

For example, if you're spending a lot of time on that particular task and it does not require a lot of accuracy, go ahead and give it to AI.

But if it requires a lot of time, but it does not it needs a lot of accuracy, then go ahead and include a human in the loop.

Right? So, understand with time and accuracy that how what are the opportunities in your company where you can go ahead and improve. For example

we have mapped the product manager's flow product development cycle into these three phases.

Now, you have to mark that where do you invest more time? What

are you not able to do?

Right? So, you are able to do competitive research.

You are able to competitive research on one competition.

But you have you don't have time to conduct it on 10 other competitors. So,

you can use AI there.

Right? There are multiple market research frameworks. SWOT analysis,

research frameworks. SWOT analysis, PESTEL framework. There are so many

PESTEL framework. There are so many frameworks that you want to use to do market research. But you do not have

market research. But you do not have enough time. Right? So, you can use AI

enough time. Right? So, you can use AI there.

Right? So, understand

what is the time taken what is the accuracy required and what is the leverage.

How this task is going to go ahead and help you.

Right? This is the second part which is called as understanding the opportunity.

Understand, guys? Everyone

everyone every company would have a different opportunity. This is where

different opportunity. This is where this framework starts helping.

Right?

Yes. Guys, understand time saved is one of the parameters, not the only parameters. Right? This is the second

parameters. Right? This is the second part.

The third part is the workflow. Now,

this is where the things get interesting.

A lot of people are using AI just at a very surface level.

You have to go actually in deep. This is

where the workflow will come.

So, what you will do is you will go ahead you will go ahead and understand exact workflows. Which is right now

workflows. Which is right now I do step number one.

I do step number two. I do step number three.

Where is the road block? How can I use AI there?

Right? Let me show you an example.

So whenever you are doing user research what are the things that you do? You

first identify right users then you reach out to users

then you conduct interviews and then you analyze the interviews take notes

and then you create persona from that profile and then you use it to make decisions.

Right? This is the current flow right now.

Right? Or maybe I can do one more thing.

Now, here there are multiple nuances, okay?

This is the workflow. Right? I will go one more I will go one more level down. Right?

When you conduct the interviews you need a set of questions.

You need a set of questions. Okay?

Digressing from a topic a bit guys, tell me if you people ask your mom that mom

whether I am fat or fit or thin what is your mom going to tell you?

Let's say you ask your mom Mom, am I fat? What is your mom going to tell you?

Provided that you have a good relationship with your mom.

Yes. So, most of the times uh because moms are one of the nicest creatures you are going to generally hear something that you love to hear.

Right? You will not hear honesty from your mom in these kind of situations, right?

But if you go ahead and stand up on top of a weighing scale of a weighing scale and if you show mom that go ahead and please read what is the number over

there. Now she cannot lie. If your

there. Now she cannot lie. If your

weight is 140 kg, then you will know what you need to do, right? If your

weight is 35 40 kg, then you know what you need to do, right? So in product management, we have something called as Mom's Test.

Which says that you should articulate your questions in such a way you should articulate your questions in such a way that the you should articulate your questions in such a way that other people will not

be able to tell lies. Okay, it's a very interesting book. Everyone should go

interesting book. Everyone should go ahead and read it. You can read it within couple of hours. Very good book.

But in order to ask questions, you should always follow the principles given in the Mom's Test book.

Okay, understand guys? Every time if you're writing these questions and all, you might get lazy.

Right? The issue is not productivity.

The issue is that we generally want to be lazy because oxygen is consumed a lot when the brain works and the brain does not want to do that, right?

So you can just go ahead, use AI in order to generate the questions for the interview with the help of Mom's Test. So you can ask AI that this is my

Test. So you can ask AI that this is my situation.

Can you create some questions for me which follow the Mom's Test principle?

Right? Which follow the Mom's Test principle. And understand guys this is

principle. And understand guys this is one use case of AI. You can use AI almost everywhere, but you have to understand the workflow. Right? So

opportunity is a bigger thing, right? I

can conduct user research, but what exactly in the user research is actually workflow.

Right? And I can go one step further.

What I can do is whatever questions that I've conducted right? Once I'm done with the interviews

right? Once I'm done with the interviews I'm going to ask AI to create 10 personas and ask AI that if these questions are asked to these 10 personas, how they are going to behave and why.

Right? So this is how the workflows happen. So workflow is you are going to

happen. So workflow is you are going to write step number 1 2 3 4 5 and then you are going to ask yourself where I can go ahead and use AI.

Right? So

we have talked about three parts. P O W, possibility, opportunity and workflow.

Everyone, a quick yes or no if you're able to understand this.

Yes.

So guys, understand in product management, we have something called as a problem space and the solution space. Okay, many people make a

solution space. Okay, many people make a mistake that they eventually or they they go ahead

and straight away jump to the solution space. Which is they start figuring out

space. Which is they start figuring out what is the tool that I should use, what is the prompt, what is the what is the exact prompt that I should go ahead and give it to it, right? And how can I get

more templates and more inspiration?

Right? This is the major mistake that people make. Understand guys? If you

people make. Understand guys? If you

truly want to become an AI native professional if you really want to follow the first principles of product management, start with POW, possibility, opportunity and workflow.

Right? This is a typical product manager mindset that first focus on the problems, focus on the opportunities before coming to the solutions.

Right? The second part, understand this, ER.

The E stands for engineering.

Now this is where Now this is where the tools, the prompts, the templates, the Claude, the lovable of the world comes.

Many people do a mistake that they start with this engineering part. That I will figure out the tool and then I'll go ahead and see what the tool can do.

Right? You have to follow the possibility, opportunity and workflow part. I can assure you no matter where

part. I can assure you no matter where you are no matter how you are working as a product manager or any professional, if you follow this framework, you should be able to become an AI native professional and you

should be able to use AI not only to save time, but to create better outcomes.

Right?

So the fourth part is engineering.

Right? And in engineering part, we have multiple things.

Okay, and this is sadly the most interesting thing for everyone. Everyone

wants to understand the tools and everything. They do not want to

everything. They do not want to understand the fundamentals which actually helps you become a better individual.

Right? Now let me go ahead and go to the the engineering part.

Okay.

Yes, you guys need to see workflows.

Just give me a moment. Let me show you if I have some workflows. Just give me a moment.

Okay, you guys need to see the workflow.

Just give me a moment. I'll show you the workflow.

Just give me a moment. I'm bringing that slide.

Cool.

I hope you people are able to see my screen right now.

Right? Where you can see a deck, a slide. Let me remove this.

Okay.

Yes. Now look at this carefully.

Right? This is what a workflow looks like. So what we do is

like. So what we do is So opportunity is I'm just giving you generic examples so that you are able to understand.

When you go ahead and conduct a research on Google it generally takes a lot of time because this is the method. This is how research happens.

You are going to ask a query on Google.

You are going to do manual understanding of all the topics, all the things. Then

you are going to type more queries. Then

you are going to write your articulation and then your output would be there.

This is how you do research on Google.

You will search something, multiple links. You will go through the links.

links. You will go through the links.

You will collect your understanding.

Then you have done the research.

Right? But now what you do is you write a query.

You think that maybe the query is not good. You are going to improve the

good. You are going to improve the query. You are going to give it to the

query. You are going to give it to the agent. The agent is going to do it and

agent. The agent is going to do it and you are going to get the output.

Perplexity Google deep research Gemini deep research, Claude deep research, all of them do the same. So

this is our example of a workflow. What

you used to do before and what you are going to do after.

Right? Another example is before Cursor how do we write code? We used to write code.

Then we are going to have some errors.

Then we are going to Google the errors.

Then when we are going to Google the errors, we are going to have some Stack Overflow community or we are going to have some documentation where it is mentioned. Then we are going to put it

mentioned. Then we are going to put it out there. We are going to debug it.

out there. We are going to debug it.

Then we are going to again change the code. Then we are going to test and then

code. Then we are going to test and then we are going to repeat.

Right now I can just give a prompt to Cursor. Cursor will find the bug. It

Cursor. Cursor will find the bug. It

will go ahead and refactor. It will run the test. It will give me the output.

the test. It will give me the output.

Right? So this is exactly the workflow.

Workflow is nothing. Step-by-step things

that you are doing right now and how you are going to improve them.

Everyone, a quick yes or no if you're able to understand workflow.

Yes. And guys, understand this takes time. Do not think that it will happen

time. Do not think that it will happen magically or you can ask ChatGPT to do this. You know your workflow better than

this. You know your workflow better than ChatGPT.

Yes, cool. Now

after the workflow, we have an important part which I was talking about.

Which is the engineering.

The sad part is that people straight away jump to engineering.

They start identifying the tools or the Geminis or the Claudes of the world rather than understanding the possibility, opportunity and the workflow. Understand guys? If you're

workflow. Understand guys? If you're

doing if you're trying to solve a problem that does not exist, you are eventually going to fail.

Right? So first try to identify the opportunities and problems before jumping to the solutions. Okay, but now once you have figured out, you have a workflow. Now you want to improve that.

workflow. Now you want to improve that.

Let me give you a set of tools. Okay.

Now there are multiple levels of the tools okay?

Level zero is ask ChatGPT Claude or Gemini.

Right? Ask ChatGPT, Claude or Gemini straight away.

For many of your problems, for example, if you want to go ahead and write an email if you want to go ahead and maybe write a press release for your product, then you can just go ahead and use the level zero part. That is more than enough,

zero part. That is more than enough, right? 80% of the use cases can be done

right? 80% of the use cases can be done in level zero. Please do not try to retrofit complex tools where they are not needed.

Right? First part.

Then we have level one.

On level one, we have things such as, for example, if there is something repeating that is coming. For example,

you want to create a PRD that should be written in a following in the particular structure.

Right? Or you want to write something that should be inspired by some kind of content.

Right? So, for that, what you can do is you can use custom GPTs.

You can use Claude projects.

You can also use something called as Gemini gems. Right? You can always use Gemini gems.

Right? You can always use Gemini gems. Right?

Tell me, guys, yes or no if you know what a Gemini gem is.

Tell me yes or no if you know a Gemini gem.

Cool. Let me show you.

Let's say Let's say I want to create Let's say I want to create Yes. Let's say I want to create

Yes. Let's say I want to create a simple product through which I'm able to generate the questions for the interview

which all follow the Mom test principle.

Okay. So, I want to create a product where I can just give the idea and it will generate all the questions for me and all the questions are going to

follow the Mom test principle. This is

the use case.

Right? So, how will I do it? First

I will go to Chat GPT or maybe I'll go to Claude.

On Claude, I'm going to ask On Claude, I'm going to ask Okay, before Claude, let me go ahead and give you one more thing, okay? Let me

show you Gemini gems first.

Okay. So, Gemini gems are small utilities through which you can create you can give custom prompts.

Okay, let me show you an example.

So if you go to this URL gemini.google.com/gems

gemini.google.com/gems

at create, you can search it on Google.

I can give a name to my gem. The name is user interview question Okay maybe maybe let's take a funny use case, okay?

Roast your resume.

Okay.

This gem will roast your resume. Okay.

Now, I can go ahead and write You are a professional resume reviewer who is also

sarcastic and brutal sometimes.

You will give a detailed roast of a resume given to you by the user.

Given to you by the user. If you want to give more information such as templates and all, you should be able to do this, okay?

Make sure you are respectable yet funny and a bit hard on the user.

Okay.

Let me save this. Everyone, tell me a quick yes or no if you're able to understand what is happening here.

Tell me yes or no if you're able to understand what is what will happen here.

Cool.

Let me put my resume.

If I have this.

Mm, there is the resume.

I've not created a resume since years.

Cool. Maybe I'll take a sample resume from Yes.

Yes. Guys, we are going to wrap up the session in the next 15 minutes, okay?

For more Yeah.

Correct. Look at this. Oh, John, where do I mean this resume looks like it was formatted on a typewriter by someone who is afraid of entering the enter key.

Right? So, now, guys, although this is a funny example understand that whatever I have shown you on this particular page whatever I have shown you on this

particular page. Where was that?

particular page. Where was that?

Almost all the problems can be solved by a gem. You just need to give

a gem. You just need to give instructions. You can I have written the

instructions. You can I have written the instruction myself, but you can generate these instructions with Claude, Gemini, or something else.

Right? And then you should be able to solve most of these problems with the help of gems. Right? Yes or no, everyone, if you're

Right? Yes or no, everyone, if you're able to understand gems. Gems are your custom mini AI apps.

Right? Now, the possibilities are limitless.

Right? You can do so many things. For

example if you want to write a PRD but you forget about what you forget about what is the format and all. You

can just ask GPT that based on all the best PRDs of the world go ahead and create a super prompt for me to create a PRD. Then copy this

prompt to chat to Gemini gem. Right? And

you can also attach two, three good examples of PRD. Now, your PRD problem is solved for life.

Right? So, this is the This is the second part. And guys, this is truly powerful. Right? And I do not want to spoon-feed you. You have to figure out what could be the different kind of use cases for gems.

Yes.

Correct. Right. So, guys, understand?

You can also do this work with level zero tools as well, but you have to repeat yourself. Here, you do not have

repeat yourself. Here, you do not have to repeat yourself.

Right?

So, Google has gems Claude has projects and they have custom GPTs, okay? Now,

the third part is Now, we have level two.

Right?

Level two is you go ahead and you build some small apps on top of these.

Okay, understand? Level one, I forgot one thing which is very viral these days, which is nothing but a collection of prompts called as skills.

Right? It is called as skills. So

a skills are a set of instructions.

A skills are a set of instruction that Claude is going to follow while doing any task for you.

Right? For example, when you are writing something with AI it is going to follow a pattern. You

will be able to understand from a very long distance that this is generated by AI. Okay. So, Claude has this skill

AI. Okay. So, Claude has this skill called humanizer skill.

Right? So, Claude has this skill called as humanizer skill.

You can use this in Claude and you can also copy this to gem.

Right? Look at what is written here.

Look at what is written here.

You are a writing editor that identifies and removes signs of AI-generated text to make writing more natural and human.

This guide is based on Wikipedia signs of AI writing page maintained by WikiProject AI clean.

Right? When given text to humanize, identify AI patterns, this, this, this, this this this.

It is going to make sure that this is done.

Right? These are Like, you can also create your custom skills and many skills are already given to Okay. There is a website called as

Okay. There is a website called as skills.sh

skills.sh where you should be able to find many of these skills.

Okay. Before installing them, read like try to ahead and understand what they are.

Like there is a skill for front-end design, there is a skill for deploying things on AI. There are

multiple skills. There are skills for AI image generation, so many skills.

Right? So, the level two is level one is custom GPTs projects gems and skills, which are nothing but repeatable instructions.

Right?

Yes.

A quick yes on everyone if you're able to understand level zero and level two level one before I come to level two.

Yes.

Now, level two is building apps.

Right? Building apps. So, here you can use lovable.

You can use Google AI Studio, where you think that you need a interface which is more mature than Gemini.

You need an interface where you are able to see multiple kind of inputs and outputs. Right? This is what you need.

outputs. Right? This is what you need.

Right? For example,

uh Where is the link?

Yes.

For example, here.

Right? You need to do follow up on meetings or you need to do analytics.

Right? Guys, can you quickly tell me a check whether I am audible, whether I am visible, whether my screen is visible?

Please tell me yes or no in the chat.

Cool. So, I am just going to take five more minutes in the session. After five

minutes, I'm going to share a feedback form with you. Anyone who shares that feedback form is going to receive a list of resources from us, free resources from us. Okay? And they will also get

from us. Okay? And they will also get high-quality recording. The recording

high-quality recording. The recording that is there on YouTube right now might not be high-quality, so you'll get the high-quality recording as well.

Okay? Just remain in for five more minutes.

After that, I'm going to share with you the feedback link.

Okay?

Cool. So, here what you can do is if you need let's say if you need to create landing pages.

Right? If you need to create apps or prototypes or experiments as a PM, right? You are not only talking about

right? You are not only talking about your own workflow, but you're talking about improving the experience of the user. Understand what happens is

user. Understand what happens is as a PM, initially, if you want to build landing pages, small experiments, it was very much technical and it would take you and

your engineers a lot of So, what happens is with different kind of tools such as lovable, AI Studio, Bolt, you can now go ahead create your

custom products. Okay? This is generally

custom products. Okay? This is generally used for This is generally used for proto typing.

Okay? People who fill the feedback form, I'm also going to send them a detailed video of how to use these prototyping tools.

Right? And the last part is agents.

Which is there are some tasks which are complex.

Right? There are some tasks which are complex, where you cannot just have a out-of-box solution that someone is going to give you.

Okay? For example,

you want to do you want to do competitive research.

Right? In competitive research, what you need to do is you need to first list the competitors.

After that, you need to go to their website.

Then you need to observe what are they selling, what is their unique selling proposition, what are they promoting, what are their features, what are their pricing, and then you have to create a

report.

Right? This is how you do user research.

Right? But understand this is how you do competitive but understand guys, this is limiting because your competition is changing every day.

Today he's your competitor, tomorrow someone else's competitor will come. In

a week, the competitors are also changing their websites and everything.

So, this cannot be a one-time activity.

You need a agent or you need a system that is going to crawl all of your competitors, identify new competitors, and it is going to do it every day.

Right? And in the morning, it is going to send you the report.

Right? This thing you cannot do with prompting with chat GPT, you cannot do with let's say other tools.

Right? A quick yes on everyone if you're able to understand the use case.

Yes, everyone a quick yes or no if you're able to understand the use case.

Right? So, understand guys, before AI, people were not even thinking about doing this because this was almost impossible.

Right? But now it is possible. How to do this? I'll show you just in a moment.

this? I'll show you just in a moment.

So, you guys are able to see Claude on my desktop right now, correct?

You guys are able to see Claude, right?

Yes. Now, look at this.

I will just go to I will just go to Claude code.

Maybe Claude code work. You can use any of them. Both have

of them. Both have six ask.

Help me build a competition tracker that does the following.

Identify the list of competitors for a given business.

Crawl the websites of these competitors.

Create a comparison chart and change the log of pricing

offering team or anything that you anything

that you find is relevant.

And then run this thing daily at 8:00 a.m.

and send me an email with this detailed report.

Right? And guys, understand my Claude is configured, so it already knows that I am a founder of HelloPM, this is what I do, and it already knows most of the stuff about me. Right? So, I will just

click on maybe act without asking.

Continue, and then I'm going I can also choose a model, but I'll just go ahead and click on this.

Right?

Yes. Everyone yes or no, tell me if you're able to understand this. Can

agents switch off my AC at 3:00 a.m. and

turn off the fan B? Yes, they can do this.

They can do this. First

First option is that your AC should have a Wi-Fi.

Second is there are physical Wi-Fi button these days. Okay? You can just put that button and then you should be able to control it.

So, I have a Mac.

So, I have a Mac Studio, which I'm able to control through a Wi-Fi enabled button if I want to go ahead and anytime want to shut it down.

Yes. So guys, understand now I am trying to show you the near the real versions or the recent versions of what Claude and chat GPT can do.

Right?

So, the tools are the same as the first version, but now we are using their agentic capabilities.

Right? So, we have one answer.

What is the business you're trying to go a different business? Maybe I'll see a different business.

Who picks a competitor list?

You research and propose them.

I'll give you the list.

How do I connect email should be sent?

Save report file each day.

What should we cover? Pricing changes,

product offering, team hiring, marketing. Okay. So guys, what are you

marketing. Okay. So guys, what are you able to understand from these questions? Tell me what are you

these questions? Tell me what are you able to understand from these questions.

The tool is Claude co-work. You can do the same in Claude chat or Claude code as well.

Yes, very good guys. Clarifying

questions very important.

This is a skill you should have as a PM.

Right, I have answered all the questions.

Then it is doing the job.

Right, so it is going to create an agent for me.

This agent is going to run. Claude very

recently, last couple of weeks only, it has released something called as Claude agents. So, you can also put this agent

agents. So, you can also put this agent on cloud. So, even if your computer is

on cloud. So, even if your computer is shut down, you should be able to your agent will keep on running.

Right, so it will go ahead and create.

Maybe I should give a business as asana.

Yes, it will do the job.

Yes, guys, this I'm not sure if you're available in the free tier, but I would request guys understand if AI is able to give you so much impact, is spending 20 more dollars would not be a big call.

Yes, cool. Now guys, tell me on a scale of 1 to 5 if you are able to understand the levels.

If you are able to understand the levels.

Right so now after that, so now we were here.

So, we have talked about the possibilities. We have talked about the

possibilities. We have talked about the opportunity workflow. Now, you also know

opportunity workflow. Now, you also know about engineering.

Right, in tomorrow's session, I'm going to give you some more examples of these engineerings. And the last part is

engineerings. And the last part is reflection. Reflection truly means

reflection. Reflection truly means whatever the tool has done, you should understand from this. You

should look at the gaps. And eventually,

you should go ahead and improvise this so that you are able to go ahead and do better with AI next time. Right, many

people miss out on this reflection.

Reflection means what did you learn? As

a human, your purpose to stay on the earth is to learn and make yourself better so that you can be more useful to other people.

Right, so reflection is the ultimate part. Now,

part. Now, I'm not talking about any tool.

You can just use any tool that you want to. And then, it should just be able to

to. And then, it should just be able to cater to these three parts, POW.

Right, reflection is done with evaluation. Yes.

evaluation. Yes.

Cool. Now guys, this was it for the today's session.

Tomorrow, I'm going to cover even more tools and even more topics. Maybe I'll

go a bit more deeper as well. Okay.

Now, I would also recommend you guys to check our program at HelloPM. So, people who are really serious, this is not for everyone.

This is only for people who are really serious about upgrading their career in product management and also in AI. Okay,

if you really want to go ahead and if you're able to commit your time, go ahead and try to look at what we have to present. Just click on this

to present. Just click on this particular video on the HelloPM website.

You should be able to understand almost everything.

Right, but understand this program requires time and effort. You'll not be able to just do it if you're not able to spend at least 10-12 hours every week.

Okay, I'll leave it to you. I'll give

you more details of the program tomorrow. Right now, my request to

tomorrow. Right now, my request to everyone is please go ahead and where is the session?

Yes, so everyone, we talked about a lot of things today.

Okay. We first talked about We first talked about what is an AI native PM?

Then I gave you the whole product development workflow.

After that, I give you understanding of the power framework. I give you my use cases. And I told you the different

cases. And I told you the different levels of how you can go ahead and implement AI.

Right, tomorrow we are going to talk about these things in a bit more detail.

And I'll also also go ahead and talk about some common questions that people might have. Okay, CVs getting rejected,

might have. Okay, CVs getting rejected, you are not able to get shortlisted, how you can become an AI PM coming from a non-technical background. I'm going to

non-technical background. I'm going to give you resources for all of these tomorrow.

Okay, but before that, please go ahead and share the feedback in the form that I have given.

Okay, I have shared a feedback form in the chat. Everyone, please go ahead,

the chat. Everyone, please go ahead, fill this feedback form so that I'm able to invite you tomorrow and I am able to give you all the resources that I have promised.

Okay, and it is also my request to everyone that if you found the value in this session, please go ahead and tell about it to other people. Share the link with other people.

Right, knowledge increases when it is shared.

When I started, I did not know how to figure these things out.

But when I start teaching, when I started trying to help more people, I was able to understand that I was the one who was learning the most.

So, sharing will give you a lot of advantage in your life.

Everyone, please fill the feedback form.

And after that, you can take a leave. We

are going to see each other tomorrow at the same time. And I will go ahead and go into deeper use cases tomorrow.

And we'll also talk about a bit about LLMs and some more case studies.

I hope everyone found the session helpful today.

Okay, please don't mention the groups.

Not good for spamming.

Yes.

Tomorrow at the same time, 8:30 p.m.

IST.

Yes.

Go guys, bring your friends tomorrow.

And also guys, if you have missed anything today, in the initial 5-10 minutes tomorrow, we are going to also revise whatever we have done today. Okay, so even if you miss something, that's okay. I'm going

to revise everything.

Cool.

Thank you guys. Bye. Take care.

Search for HelloPM on YouTube. Go to the live section. You should be able to find

live section. You should be able to find the videos.

Yes, I'll not be sharing interview hacks and resume hacks, but I'll share resources which I have created that will help you all guys.

I have shared the feedback form once again in the chat. You guys can go through this.

Thank you everyone. Bye and take care.

You have been an awesome audience. I

believe you'll be able to go ahead and