Build AI Simulations in Storyline in 10 Minutes | Live Replay

Welcome everyone! Glad to see you here. I am very excited for

you here. I am very excited for this session. I have been

this session. I have been working very hard on it for the past couple of weeks. Very

excited to be here in this moment. So thank you all for

moment. So thank you all for being here with me.

Today is the official public beta launch of this new simulation builder tool that I've been working on day and night it feels like. So I can't wait to actually share it with you all, see what you create.

Are you all familiar with the tool already? Have some of you

tool already? Have some of you all been seeing little sneak peeks on LinkedIn? Or some of you all hear no idea what we're about to get into.

Because I'm most excited for you all. You're going to be

you all. You're going to be surprised I hope. No idea. Okay

great. Okay very cool. So I

hope that you all will be surprised here. I try to make

surprised here. I try to make this as easy as possible for you all.

I have a question before we dive in though. So for those of you who do design scenario based learning experiences or eLearning simulations or whatever we want to call them. What has been the hardest part so far when you work on those?

So it looks like the... It's

honestly pretty close. It looks

like the hardest part for most people is the branching complexity and the time it takes to build and making it feel realistic are very close seconds.

So I know the struggle. I know

one of my first eLearning projects I did for a client back in like 2018. I was like oh I'll do a branching scenario.

I'm seeing people talk about it on LinkedIn. Seems super

on LinkedIn. Seems super excited. I'll make it this yeah

excited. I'll make it this yeah really cool scenario based simulation style eLearning.

And I won't even storyboard right. I like rapid prototyping

right. I like rapid prototyping so I'm just going to dive in and start building it in storyline and that course must have taken five times longer than I expected it to with all the different branches and things that could go wrong.

So yeah it takes a long time.

The branches get complex. So

thanks for sharing that. It was

more of a spread than I expected. So I'm going to share

expected. So I'm going to share my screen and we are going to get into the good stuff here.

Okay I hope you all are excited.

And we have some people who will be helping out in the chat.

So if you have questions in the chat feel free. I'm going to maybe stay focused for a few minutes at a time.

So here we go and make sure all my systems are working. It

looks like they are. Okay so

the traditional approach to building simulations in storyline. It takes a lot of

storyline. It takes a lot of time right at least weeks sometimes months if you want to build it really well.

And most of the time we're building multiple choice experiences right. Especially

experiences right. Especially before AI. We didn't have the

before AI. We didn't have the option to use AI powered characters or things like that.

So you know in the real world people aren't going to have the option when they're talking to an angry customer like oh which choice do I pick to calm down this angry customer.

So it's like it was the best thing we had to help people practice this decision making in a self paced environment.

But it still was not the most realistic because people on the job don't have little multiple choice options floating around their head when they're having to make a tough decision in most cases.

And then also the visibility into the learner performance is not very great right. We could

sometimes rig a score onto one of these branching experiences where they get points based on the outcome they arrive at. But

we can't really see the exact path they took the exact skills they're struggling with unless we spend a lot of development work on implementing something like XAPI for example. So yeah

not to mention the technical skill it takes to build build simulations like this. So I

built this tool to make it take literally minutes.

And I'm not even exaggerating here you can build this in less than an hour. So yeah you can build these realistic simulations really fast learners respond in their own words and get dynamic responses.

So these are AI powered simulations so no two people's experiences are going to look the same.

And right no matter how you talk to the AI it can respond as as the person that you're coaching would or as the customer would that your learners need to support or that the you know if you're training salespeople that the prospects would respond so we can make a simulation that responds realistically like people would in the real world.

And then we can track every conversation too so we can see the exact criteria that people are struggling with in these simulations we can see full transcripts we can see score pass fail all of that good stuff as well.

So do you all want to see the simulation we're going to dive into here. You want to

into here. You want to interview an AI powered subject matter expert.

Okay let's do it let's do it so I'm going to share this link with you all in a minute. Here's

the demo okay we're going to we're going to interview this guy. He

retires in two weeks.

We need to collect the right info from him before he leaves so that we can design a high impact learning simulation for new managers, we should be able to do that right.

Let's do it.

Okay, so here is where the simulation loads.

And this is the whole AI powered experience right here.

So Frank says they've got me doing a knowledge transfer thing two weeks left and now we're capturing my institutional knowledge, no offense.

But when someone did this with Diane and claims, it became a PowerPoint nobody opened. What

exactly are we planning to build.

So let's just be a little vague so I'll be like we're going to build any learning experience.

So Frank is thinking here. Oh,

Frank did not like that.

So Frank groans.

He says oh an e-learning experience great last year they made me do a 45 minute compliance module 45 minutes. I

clicked the next as fast as I could.

Is that what we're building here because if so, if so I've got a stack of end of year reports that are honestly more exciting.

Okay, he's not too excited about that. So I'll be like, oh

about that. So I'll be like, oh no no no Frank.

We're going to build, build something much more engaging.

It's going to be a scenario based simulation, where people can practice, practice using the same skills they'd use on the job.

Okay, Frank like that. So it

knows we got a little star right.

Add a little gamification element where if we can confirm that we are going to build something a bit more engaging.

Frank, Frank gets happier. So

Frank says, okay, that's actually different.

If it's a real scenario where someone has to practice doing the actual job, make real decisions, maybe feel that not in the stomach when an agent pushes back on feedback.

Yeah, I'd actually use something like that. All right,

you've got my attention. What

do you need from me?

So how do we want to start this conversation? Let's start. Let's

conversation? Let's start. Let's

ask a fluffy question and see how Frank responds.

So I'll be like, what does it take to succeed in your role?

All right, let's be a little big and you all will all have a chance to try this for yourself.

I'd love to see how y'all handle this conversation.

Okay, Frank's given us that look. I mean, you got to know

look. I mean, you got to know your people, handle the tough situations, keep the numbers up, you know, the usual stuff.

Okay, so this me, I guess doesn't love the fluff as much as some SMEs might. So we are going to

SMEs might. So we are going to have to take a different approach here.

So let's try to get more specific. Actually, we could

specific. Actually, we could take a wrong turn.

Y'all want to take a wrong turn and be like, okay, know your people, handle the tough situations, keep the numbers up.

Sounds good.

Be like, sounds good. Maybe we

can wrap this conversation up.

We got what we needed in this one minute conversation, right?

Oh, Frank is alarmed again.

That's it. You don't want to know about the hard part. That

generic stuff. I just said it.

Any manager on the street could tell you that. Okay, you're

right.

Tell me about the hardest situation you've dealt with.

So now I'm going to try to, you know, finish this up and actually get to the feedback, right?

So now Frank's thinking we might be on the right path here.

That's a little better, but still pretty broad.

There's hard situations with agents, hard situations with customers, right? With your own

customers, right? With your own manager breathing down your neck about average handle time while you're trying to actually develop somebody.

Which one are you asking about?

Let's dive into coaching the agents.

What's the hardest situation you faced with an agent?

Because this is a call center manager.

And we're presumably going to be designing a scenario to help other call center managers.

All right, coaching situations.

Hardest ones are always when the agent knows they're struggling, but they're too proud to admit it.

So they come in defensive before you even said a word.

How did you handle it differently compared to new managers?

So we'll see what Frank is doing to perform better than the other managers. Give me a second. I want to get this

second. I want to get this right.

The thing most new managers do is they walk in with their call stats print out and they lead with the numbers.

Second, you do that with a defensive agent, they shut down completely. So how do you

completely. So how do you handle it?

So, yeah, we're actually having a little smee conversation here trying to, oh, we might have finished it here.

We got another star for unlocking a specific scenario example. Notice that the

example. Notice that the simulation is complete and the results are loading now.

So this is where the evaluation is happening. Frank is happy.

is happening. Frank is happy.

And let's see, what was his actual thing?

I never let the agency my data first. I asked them how the

first. I asked them how the week went. I shut up and

week went. I shut up and actually listen.

Nine times out of 10, a struggling agent knows where they messed up. If you let them say it, they own it instead of defending against it.

So there we go. We'll continue

along. And here is the feedback.

So we failed. I guess it didn't like our trolling.

So we failed up here. We only

got a 56 percent and we get some customized feedback here, too.

So we recovered well after Frank's skepticism by pivoting to scenario based practice, and that kept him engaged.

Your questions stay too broad throughout. What does it take

throughout. What does it take to succeed in your role?

And even what's the hardest situation you faced gave Frank room to stay vague, anchoring to a specific moment.

Right. So on and so forth. So

there we go. What do you all think?

Do you all think this might be pretty difficult to build? And

you want to see the results?

So check this out. And yes,

speech, the text and voice motor both coming soon.

If you all are thinking about how this would look voice, we already have some working prototypes of both of those.

OK, so check this out. You all

are all going to build this literally during this session if you think this is going to be like too hard.

So if you don't know where to begin, you're out here in the right place. So check this out,

right place. So check this out, too.

So this is the metrics screen.

So whenever you create one of these simulations, watch, we can refresh this.

We'll see our completions come in. I did a couple. I did a

in. I did a couple. I did a test one.

We can see pass rate completion rate of how many unique learners completed it, score distribution.

And then we can actually see criteria breakdown. So I'm

criteria breakdown. So I'm going to show you all a little bit here.

I don't want to overwhelm you, but this is the evaluation criteria. Right.

criteria. Right.

So we're evaluating whether or not the learner asked a specific behavioral question that references a concrete situation.

So on and so forth. We can

assign points to each of these.

We can also account for deductions.

So if we tell Frank we're designing generic e-learning, we're going to remove 10 points.

And then in the metrics tab, we can scroll down, see our actual score right here.

We can click on it and we can see the full breakdown here so we can see how the learner did with each criteria.

Right. Did they get full credit or partial? And then we can

or partial? And then we can also see the full transcript.

So, yeah, so pretty good visibility here. And again,

visibility here. And again, this is the first day of the public beta.

So I'm going to be working on giving even deeper metrics and insights.

The simulation. So this is what powers the whole experience.

Right.

So it's a lot of text we're looking at. Right. We can give

looking at. Right. We can give the character instructions, a description, behavior triggers.

So right when the learner asks a vague question, give a vague, unhelpful answer with a shrug.

So you can see some of that here. So there's all that. OK.

here. So there's all that. OK.

I'm going to I'm going to show you how easy this is to make, though. OK.

though. OK.

Don't get overwhelmed. We don't

have to fill out all this information before we do that.

I want to show you one more example just to show you what these different experiences can look like.

So Sabrina Gonzalez created a really impressive portfolio project.

Honestly, the chat bot is like the least impressive part of the whole thing.

So check this out if you haven't already.

But we are a detective and we have to interview some suspects about a stolen painting.

So check it out. So here's how you can like brand the simulation to match different course skins and themes and stuff like that.

Yes. Did you commit the crime?

I don't think I'm going to make the best detective here, but no, can we move this along?

I've got places to be. Admit it.

So you can check this out, too.

We'll also share the link to this.

But this is a really, really impressive demo if you're looking for creating portfolio projects.

So what do you all think? I'm

going to stop sharing my screen, check in on the chat and see how we're doing.

OK, so you all love it. You all

you all feeling confident you can create one in the next 45 minutes or so?

Some of you all are curious how to do it. We're going to dive into the demo really soon.

I do want to play a quick clip for you from some early beta tester.

So some people are using it for their portfolio.

Some people are using it on the job for their audiences and figuring out how to make it work.

So I'm going to play a four and a half minute clip.

These people were kind enough to record some clips talking about their use cases for us.

And I thought it would go smoother as a video than having to manage it all live.

So four and a half minute clip.

Then we're going to dive into the demo and you will start building.

So get storyline open and get ready to go if you want to follow along.

Everyone, my name is Patty Blount.

I am a senior instructional designer for Optimum.

We are a telecom services provider from Bethpage, New York.

My role is predominantly sales training for our entire organization.

So that means multiple channels, multiple business units such as consumer as well as business sales.

I wanted to share with you how the Devlin AI is solving a few problems that we've been having.

First, we are a very tiny team with no developer resources.

Every time we explored AI in the past, we always hit the same problems, the same brick wall.

How are we going to handle JavaScripting? Who's going to

JavaScripting? Who's going to manage the API keys?

How will we manage maintenance going forward when we have to update these?

That technical lift kept us from ever really getting started.

Second, we need assessments. We

need a way of delivering assessments so that we can actually scale our programs more efficiently without losing consistency.

Our assessments have to be role play based because multiple choice simply does not measure the performance we are looking for.

We tried and we are still using train the trainer sessions with those strict rubrics I mentioned.

Again, the more people that you introduce to the equation, the more consistency goes down.

Devlin's AI addressed both of these issues.

I was able to build two simulations using the beta version and maybe about an hour.

It required no coding, no API keys to manage.

I was able to just simply tell the AI how I wanted it to behave.

It provided to me all the storyline variables I needed to add to my project pile and a few JavaScript lines that I had to embed in a variable.

That was it. It's all on the same slide. Very small, compact,

same slide. Very small, compact, and very elegant solution.

Devlin's AI didn't just save me time.

It made my assessments more consistent, more scalable, more genuinely useful to my learners.

So if any of these technical barriers have been holding you back, I urge you to consider taking a close look at this simulation builder.

It's well worth your time.

Hello, my name is Terry Arthur and I'm an instructional designer with Yamaha Golf Car.

I'm a team of one, so time and simplicity are very important to me.

We currently have use cases where role play or skill practice would be valuable, but I just don't have the bandwidth to make it happen.

Being a beta tester with Devlin Simulator has been a game changer for how I can now create the interactive content that will have a real impact on learner behavior.

I've created three Sims, each taking me less than 30 minutes to build and load into storyline.

Thank you, Devlin, for creating this amazing tool.

Hello, I'm Sabrina Gonzalez, one of the instructional designers at Peck Academy, and in the Academy, I've reviewed lots of projects and portfolios.

And the projects that always stand out the most to me are the ones that move beyond traditional interactions that you would expect from e-learning.

And I think that's where these chatbot style simulations fit in.

And they make the chatbot simulations make really good portfolio pieces because you do still have to be a strong instructional designer to create them.

So three of the big pieces that I think are good to show off with these chatbot styles are first that the learning experiences are much more authentic.

You can create a really solid scenario based e-learning project, but you're still limited to just what answer choices learners can choose from.

And AI opens up a whole new world because people can put in exactly what they would say.

Maybe the three answer choices in the e-learning aren't what they would say, but with AI, they can put in whatever they would say or do in the situation.

So much more authentic.

The other piece is you can't rely on AI to do everything for you. So you still have to be

you. So you still have to be the strong instructional designer who knows exactly what success is going to look like for this learning experience.

Because you have to tell AI, these are the things learners must do in order to pass this simulation.

So it's still your job to dig in and figure out what does success look like.

And then last is the feedback.

In a traditional e-learning, you can give some incorrect or correct feedback.

It's going to be relevant to those answers that they chose in the experience.

But you can make it relevant to the learner by using AI because the AI is going to give feedback based on the user's input and the success criteria that you set.

So it's much more relevant to them and much more likely to change behaviors because it's tailored to the learner.

And so they'll know what they need to do better next time.

So it's those three pieces that I really feel like make these strong portfolio pieces.

Authentic learning experiences.

You still need to know what success looks like and the feedback is going to be much more relevant to people.

All right. Big shout out to Patty, Terry and Sabrina. Thank

you all for recording those clips.

They're all working on cool stuff. And then Sabrina says

stuff. And then Sabrina says that that little detective when we took a look at.

So, again, check that out if you haven't already. So who is ready to build?

We're going to do the follow along and we're going to build this thing now.

So it is time. Let's see how to how to actually do this thing.

OK, so first things first, we need to go ahead and open the guide.

OK, so I actually created a full text based guide for you that lists everything step by step that we're going to go through during this this workshop.

So if you do fall behind or get stuck at any point, that guy that guy will be there for you after the session as well as the recording.

So I just shared the link to the guide. You can click on

the guide. You can click on that open guide button if you're here live.

And that has all the links you'll need and the instructions you'll need.

So you will be able to do this.

I promise you that I am going to hop over into development mode now.

We're going to start building this thing. So this is the

this thing. So this is the guide that you all should have been able to open.

It's just a Google Doc. This

link right here links to that interview that I showed you at the beginning of this session.

If you want to play around with that on your own time, mine has a couple more features like those stars that appear that aren't in the version we're going to create.

But once you see how this works, you should be able to add that feature relatively easily.

Second most important thing is download the storyline template.

So I'm sharing a template with you all to start from so that we can only work on basically the hard parts.

So download this template here and we're going to get started shortly on that template.

Notice I mentioned that if you'd like to use the same font I used in the template, you can download the Google font.

I don't recommend doing that right now. It's going to take a

right now. It's going to take a little bit longer. Just do that after the session.

Your template just might look a little different than mine if you don't have this font installed.

And then after that, go ahead and sign into the Sim Builder.

So you can go to devlin.ai/login

in your address bar and sign in with your Google account if you have one.

That's what I do. Otherwise,

you can get a login code sent to your email address.

So we've got a couple of things going on. We're opening the doc.

going on. We're opening the doc.

We are downloading the storyline template and we are logging into devlin.ai.

It's free for everyone to sign in. You get 5000 credits right

in. You get 5000 credits right now, which is enough for probably at least 100 AI powered conversations, if not more.

How are we doing with that? Are

you all ready?

Once you sign in for the first time to the Sim Builder, it will ask you a few questions like your name and stuff like that.

And then it will drop you right on the new simulation screen.

If you already had an account, you can click on this new simulation button to get to this new simulation screen.

If you are here, though, after signing in, you are in the right spot.

As long as you'll have the template open, which looks something like this.

And you're ready in your account.

OK, I'm going to move on because I see that some of you already remember everything that you need will be available in the replay.

OK, so first order of business will be in the devlin.ai tool.

And we can look at step three right here. This is essentially

right here. This is essentially a copy and paste job.

So we're going to copy the simulation name. We're going to

simulation name. We're going to call it the retiring Smee, the learner.

We're just going to put instructional designers.

And here is the simulation description. So you can be as nondescriptive

description. So you can be as nondescriptive or as descriptive as you'd like here.

I shared a prompt that should get us pretty close to what I used for the demo I shared.

My prompt looked a little bit different because I was like, I'm hosting a workshop around this and it needs to be like really good.

And like I need to be able to push Frank's buttons really easily and so on and so forth.

So I added a few like live event specific things to this.

But in case you didn't know where we're going with this yet, we are going to.

A.I. is going to actually create the whole simulation for us so we can describe it in as much detail as we'd like.

Right. Once you've got everything copied in here.

Click on that generate SIM button. OK, so copy in those

button. OK, so copy in those three things.

Make sure you get both paragraphs from the simulation description and then click on generate SIM and just let it do its work.

So it can take around 40 seconds for this step.

So just watch the glowing orb really, really commune with this orb here.

And before you know it, the SIM will be created.

And yes, the SIM description will be used to fill out all of the details for this simulation.

So I have a really in-depth A.I.

prompt running right now that knows which content will power this simulation the best and knows how to fill out all of our fields really well.

So it will it will fill out everything for us.

So simulation ready. It gives

us a screen where we can actually try the simulation right here.

The character is already spun up and talking to us.

So this me is saying, so they finally sent someone I figured HR would want to do some kind of exit interview, but an instructional designer is a new one.

What exactly are we trying to accomplish here? And we can

accomplish here? And we can test it out.

We are going to make an info dump e-learning course.

Kelly, the quality of the SIM should be high, regardless of how detailed your SIM description is.

So don't feel too much pressure around that.

I tried it. I've made some SIMs for the description.

I say just make something really engaging and it comes up with a pretty engaging SIM.

So there we go. So this works.

Now we can click on this continue to editor button and you can see all the magic that happened during that 40 seconds.

OK, so on the tab now we can see the simulation name, the character name.

Look at all of this, right? We

didn't write all of these instructions.

Start neutral and a little guarded. Short answers, mild

guarded. Short answers, mild skepticism.

We have the characters opening line, scenario context, learner role with more detail and look at all of these behavior triggers.

So when the learner asks a vague question, Shrug can give a general answer.

So again, the AI generated all of this from the description that we shared with it, as well as the simulation endings.

There's max turns that we could set.

So you as the designer have full control over changing or updating any of these things as needed.

But just see how much work this AI engine does behind the scenes.

It generates all the evaluation criteria for us.

Deductions. Right.

And also the best part, the storyline publishing settings.

So it generates this variable for us, tells us when the variable should change, has these different values set up, so on and so forth.

So how are we doing? Is every

are most of us at this point?

Do we have the do we have the simulation created because it's pretty much ready to drop into storyline now that it's created.

We don't need to change anything. That's all we needed

anything. That's all we needed to do to create the simulation.

And the key reason this worked is because of this line right here.

The SIM must use a storyline variable called Frank mood.

It would have probably created that even if we didn't include this line.

I just wanted to be sure since we have so many people here today.

And I said, with these available states, neutral, happy, thinking, confused and alarmed.

And if you haven't figured it out yet, the reason I said that is because in our demo file, if we go to this slide one point to the simulation side with Frank and click on him, go to his states.

We can see that we've matched up those states exactly.

So if you're new to storyline, right, we did that. Did this

just by inserting a character, editing the states and then clicking this button to add new states as needed and then customizing them as we want.

So this one point to slide is where we are going to be doing our programming next.

Yeah, we and then the questions about other languages, we could definitely make this make this work in up to 70 different languages, I think.

So yeah, definitely can. If you

do all of your settings in French, the simulation will likely work in French, even though the editor is in the UI for you as the designer is in English, but I could update that too.

We will have time for Q&A at the end as well.

But are you all ready to continue? Do you want me to

continue? Do you want me to show you how to wire it up to storyline now that the simulation is complete?

Okay, let's keep moving. So

here is, let's just follow along with the simulation guide so it makes it easy for people to like catch up if or as needed.

Or yeah, with the with the implementation guide. So step

implementation guide. So step four is embed the simulation.

Okay, so in the Devlin AI tab, we are going to go to the publishing tab, we're going to scroll down inside of the storyline 360 section.

We are going to scroll down until we get to step one in the embedded implementation guide here and we're going to copy this URL.

This is the web object URL.

Once we have that, we're going to hop back over to storyline, go to the insert tab while we're on slide 1.2 and click on web object right here at the top.

So click on web object, we're going to replace the URL that's in the address already, and we're going to paste the URL we just copied.

So it should look like this and we will click on OK. Once we've

done that, let's just resize it so it looks good over here in this space.

Right, try to center it with the continue button so it looks nice.

There we go.

Okay, so now the simulation works inside of storyline. If

you publish this right now, this simulation would work. The

only thing it can't do yet is change the states of the character.

So that's what we need to work on next. And yeah, of course

on next. And yeah, of course the character can be changed.

You can use completely custom characters yeah.

But yeah, we'll have lots of time for questions at the end.

I'll stay a little past 60 minutes to make sure I answer all of your questions.

Okay, so step two, once we've done the embed, is to add the storyline variables. So this

storyline variables. So this again is a copy and paste job.

I actually made it even easier in the implementation guide.

Let's see.

Oh, I added a step here to review the character states. We

already did that. I just wanted to make sure you know, like, the reason this is working is because we added additional states for this character.

So I already did that in the template since that's a relatively basic storyline skill.

So then we're going to create the storyline variables. So

these are the variables we need.

Again, I made it pretty easy here inside of the implementation guide, but you can also find this information right here.

The only thing is it might include some variables we aren't actually using in this demo, right? The AI created these two.

right? The AI created these two.

We're not going to need this one.

So if you want to just keep it really easy, just copy the variables that I have in the text-based guide over here.

So in storyline, we are going to click on the manage project variables button over here on the right.

And then I'm going to click on this green plus button to start creating these variables. Now,

a couple of things when creating variables, we need to make sure the capitalization matches exactly.

So Frank mood. Yeah, if it was lowercase like this, the whole thing would not work.

So we need to make sure that the capitalization is exact and we need to make sure that the type is exact as well.

So Frank mood is a text variable. We will leave the

variable. We will leave the default value, whatever it is.

We're going to create another one now.

This one will be sim eval complete. If you want to make

complete. If you want to make sure you get it exactly right, you could just write, copy it and paste it.

And sim eval complete is a true false variable. So we'll do

false variable. So we'll do true false for that default should always be false for the true false variables.

I'm just going to speed this along. Now I'm going to copy

along. Now I'm going to copy and paste that one. I will do name this name this one sim pass, since that's the other true false variable.

I'll do sim score now and don't miss these underscores that we're using. Sim score is a number

using. Sim score is a number variable equal to zero by default.

And then sim feedback is a text variable. And I think those are

variable. And I think those are all the ones we need.

OK, so how are we doing? We

need these five variables. One,

two, three, four, five. Once

you've created those five variables, we're going to create the triggers to make this this whole simulation work.

But if you all are already experienced storyline users, this should be really, really basic stuff. Like we are not

basic stuff. Like we are not doing anything crazy here.

So we've created these variables. Once I get some

variables. Once I get some confirmation in the chat that you already, I'm going to go ahead and move along.

OK, now we're going to create five triggers. These triggers

five triggers. These triggers are going to be, again, pretty intuitive. If you all already

intuitive. If you all already work in storyline, we're going to click this new trigger button right here.

And we are going to change the state of Frank to happy when the variable changes Frank mood if Frank mood is equal to happy.

And this is going to be lowercase.

OK, for those of you who are newer to this, we're basically saying, yeah, I'm changed Frank's state.

We already saw his five different states when the Frank mood variable changes, because the AI will automatically change the Frank mood variable based on Frank's mood as determined by the AI.

That's how this is working. So

we were just building the bridge here. We're saying here's

bridge here. We're saying here's what we want to happen in the actual course when the AI changes the Frank mood variable.

And so then, since the Frank mood variable can be equal to multiple things, which we can see on the publishing tab in storyline 360 settings, here is the logic for this Frank mood variable.

Right. We explain when it should change. And here are its

should change. And here are its potential values. So notice

potential values. So notice that this value is a lowercase happy.

We need to match that exactly in our condition, because that's what the AI is going to set the value to.

So I'm going to press OK to create this. And we basically

create this. And we basically just need to copy this five times. So I'm going to press

times. So I'm going to press the copy button.

I'm going to paste and then I'm going to update the one that I just pasted. We're going to

just pasted. We're going to change the state to thinking.

When Frank mood changes, if Frank mood is equal to thinking, all lowercase. I'm going to

all lowercase. I'm going to paste it again, open it up again, and I'm just going down the list, right?

We can do neutral. So I'll do change state of Frank to normal, which is the default state. We

have them as neutral by default.

When Frank mood changes, if Frank mood is equal to neutral.

I'm going to paste it again. We

should only have a couple more now.

We need to do confused and alarmed.

So change his mood to confuse or change the state to confuse when Frank mood changes, if Frank mood is equal to confused.

And then I'll paste it one more time, open it up, and that will set the state to alarmed when the variable changes, if the variable is equal to alarmed.

Hopefully this is pretty self-explanatory.

If you're new to working in storyline in this way, this just go a little slower with the implementation guide in the replay.

And this should make sense. But

again, high level, the AI is changing the storyline variables.

We are basically creating triggers to change stuff in storyline when the AI changes the storyline variables. I hope

that makes sense.

If there's any confusion, let me know during the Q&A. OK,

there's a one final trigger we need to add. Right.

Are you all ready to get into the dreaded JavaScript? Right.

The dreaded JavaScript. Here's

what we're getting into, y'all.

Oh, wait, no, there's actually this. There's this first. OK,

this. There's this first. OK,

we're about to get into the dreaded JavaScript.

So what we need to do is look at this continue button. Right.

So this continue button is in the hidden state by default because we don't want people to be able to continue until they complete the simulation and not just complete the simulation, but get their results back from the simulation. Because on the

the simulation. Because on the next slide, we show all these results.

If they haven't actually completed the simulation, they're not going to, you know, there's if they haven't even gotten the results back and they continue, this slide is just going to look empty.

So we need to create a new trigger where we change the state of the continue button to normal when a variable changes.

Sim eval complete, because when the evaluation is complete from the simulation, it will change this storyline variable to true.

So you could add a condition and say, like, you know, when Sim eval complete is equal to true, that's not really necessary because it's false by default.

The only way it can change is to true. So I'm going to click

to true. So I'm going to click OK so that this continue button actually changes.

And now we are getting into the dreaded JavaScript. OK.

dreaded JavaScript. OK.

This line of JavaScript right here is what you need to copy and paste. This is the dreaded

and paste. This is the dreaded JavaScript.

So copy that JavaScript, hop back over into storyline. We're

going to do an execute JavaScript trigger.

Click on JavaScript right here and then paste in that single line of dreaded JavaScript.

And you're done with the JavaScript. Actually, there

JavaScript. Actually, there might be another little tiny part to do.

So click OK. We're going to execute this when the timeline starts on this slide.

OK, so this JavaScript just initially initializes the listener so that the web object here has the ability to update the storyline variables.

So this this JavaScript is basically the bridge between our web object and the storyline course itself.

So this is needed for it to update the storyline variables.

So we'll click OK.

And that's that's all there is to it. So if we were to publish

to it. So if we were to publish this right now, this would work.

You want to try it. And also,

this is really important to clarify.

You will never have to do this again inside of this simulation.

So if you go back over here and you're like, I want to make some changes, right, like I want to change how it works.

I'm going to I'm going to update his character a little bit. I'm going to change his

bit. I'm going to change his first message.

You know, I'm also going to change the whole visual design of the whole thing.

I'm going to change all the colors. I'm going to change the

colors. I'm going to change the font weight.

I'm going to change the evaluation criteria. You click

evaluation criteria. You click Save Changes and within 60 seconds, those changes are going to be live in your course.

So no republishing and storyline, no updating JavaScript lines like you're already done.

So once this is live and published, you can update the simulation only from this tool without having to ever republish or do anything from storyline.

Yeah, I got to try to make it as easy as possible. So let's

see if this works. OK, this

this should work fine now.

So I'm going to publish this.

And when we publish, we'll click on Publish.

We will click on Web and this will just publish it locally.

So I'm going to publish it locally.

We'll view the project and let's see if we can talk to Frank here.

So they finally sent someone.

What are we trying to accomplish here? Well, I want

accomplish here? Well, I want to make an e-learning course.

OK, and Frank is alarmed. So if

you were able to follow along until this point, we built the simulation.

That's that's how difficult it is. So how are you all feeling

is. So how are you all feeling for those of you know, some of you all might have gotten slowed down along the way.

Again, I'm confident with this guide. You'll be able to get

guide. You'll be able to get there today for sure. But yeah,

we did it.

Was it easier than you expected?

And we did all of this just to remind you all from this prompt right here.

Right. This is all this is the level of description we needed, which was honestly even more in depth than we actually needed.

When I first played in this session, I was testing with only this paragraph right here.

So it does work and rise again.

Yeah, I'm excited to get to the Q&A for you all. I'm going to answer some questions.

I have some more slides for you.

Do you all want to build the rest of this really quick or should we move on?

Because the next piece is basically adding some variable some variable references so that we can display the feedback on the slide.

You know, you would use the I'll show you for this one like you use this percentage sign syntax and then you type in the name of the variable and then enclose it with another percentage sign.

Click out. And now when they get to this slide, it will show that simulation feedback. So

you all do you want me to build this piece out?

OK, let's do it. I'll try to do it pretty quickly. I'll do the percentage sign SIM underscore score.

And these have to be spelled and capitalized exactly like they are in the variable manager. Again, I put those

manager. Again, I put those instructions in the full guide.

We have some pass fail text here. It is fail by default and

here. It is fail by default and pass as an additional state. So

I'm going to create a new trigger.

I'm going to change the state of the pass fail text to pass when the timeline starts on this slide.

If SIM pass is equal to true.

Yes, this is this is also included in the implementation guide, so you don't need to follow along with this right now if you don't want to.

OK, so this is the trigger for that. So then when the timeline

that. So then when the timeline starts here, if our if our simulation set SIM pass to true, it will update the state of that text box to pass and it will show the appropriate thing.

The only other thing we need to do here, this is the only other line of JavaScript I was talking about, is we want to we don't want you to manually have to like change all of your storyline variables back whenever someone clicks the try again button.

So I made it really easy in a single line of JavaScript. So

click on this try again button.

Click on this new trigger. We're

going to execute some JavaScript and the JavaScript will be ready.

Our second super hard line of JavaScript. We're going to

JavaScript. We're going to spell out restart simulation and then put two parentheses to closing parentheses.

So notice the capitalization is important. Lowercase restart

important. Lowercase restart capital S simulation. There we

go. Click OK. And we're going to do this when the user clicks the try again button.

And that's all. So that will reset all of the variables associated with the simulation as well as the simulation itself.

And you can move this trigger above what we need to move this trigger above the trigger that actually jumps to the other slide or else this JavaScript will never run and it will never reset.

So I'm going to click on that JavaScript trigger and click on the up arrow so that first it will execute that restart JavaScript and then it will jump us back to the simulation.

And these instructions, if you're new to this kind of thing, these live in the full implementation guide within the storyline 360 publishing section as well.

So all the way at the bottom here, here's where it gives the instructions on restarting the simulation. It's included in

simulation. It's included in the implementation guide as well.

So I think we're just about done here. That's the results

done here. That's the results slide. Did we miss anything? We

slide. Did we miss anything? We

did that. We did this. And

there we go. So now you can just publish and test.

So that's all there is to it.

Again, you can, you know, look at your specific instructions.

Again, they were all AI generated. So your stuff might

generated. So your stuff might look a little bit different.

You feel free to refine this one. But of course, what I'm

one. But of course, what I'm most excited about is seeing what you all create. This will

work in storyline three. Yes.

I mean, I haven't tried it, but there's no reason I can see why it wouldn't work in storyline three, I guess is what I should say.

OK. Let's see. Let's see. I'm

going to share a few more slides with you all. Then let's

dive into the questions. I can't

wait to hear what you all are thinking.

So this is probably a question on a lot of your minds, right?

The security, the privacy. How

can you use this at your actual organization?

So I am designing this for enterprise use from the start.

So this isn't just like a nighttime hobby project for me.

Like I once I shared some initial prototypes and saw how much excitement there was on the enterprise front.

I'm going all in on this front.

So I'm making sure that we can roll this out to actual audience of audiences with thousands or tens of thousands or more learners at a time.

And I'm making sure we can comply with enterprise requirements. So learner

requirements. So learner conversations are never used to train AI models.

So this isn't like a data harvesting operation. Your data

harvesting operation. Your data is encrypted when it's moving between endpoints and at rest.

So your data is safe. Tenant

isolation via row level security. So in our database,

security. So in our database, user A physically cannot access data from user B, you know, hard separation at the database level.

We're using Versel and SuperBase for hosting. So this is

for hosting. So this is enterprise great architecture here. This is all hosted in the

here. This is all hosted in the United States if you needed to know that info.

And for those of you who are in the know about SOC 2, we are beginning this compliance process, which is the most industry standard enterprise security, privacy data, data handling process.

This is essentially an independent audit of our entire system. So, yeah, we're

system. So, yeah, we're designing this for enterprise pricing.

I've seen some questions about that in the chat. So I have this generous free tier, giving you all some credits. Very good

for hobbyists, for portfolio users.

After our beta is over, it will be $49 per month per designer.

During this founder's discount period, though, it will be $39 a month for life.

This includes 35K credits per month. So for three more weeks,

month. So for three more weeks, we're going to be running this founder's discount just for all the people who are believing in us early.

I do appreciate all the support for those of you who are believing in us early. I've

been working incredibly hard on this, so I really can't wait to see it start rolling out to bigger audiences and seeing it on portfolios and things like that.

So that's the pricing. But

again, you can see if you click on the credits up in the top right, if you click on this, it will show you like what's included in the different tiers.

I'll probably add a different button or a banner or something like that. But yeah, you can

like that. But yeah, you can get like the score of my next API export. You get those 35K

API export. You get those 35K credits a month.

You can turn off all the Devlin AI branding from your widgets, update unlock basically unlock the core platform. But

the portfolio tier you can do for free as long as you're okay with the branding.

Also, two new modes coming soon.

I wanted to have these demos ready for you here, but I was really trying to make sure, you know, the core platform was as strong as possible before the public beta.

But we do have working prototypes of a coach mode. So

essentially what that looks like is while you are talking to Frank, you could essentially access like a coach and then a coach AI powered dialogue comes up.

And then that coach can see the full transcript, the full conversation you're having with the character and the coach can give some guided support while you're also talking to the character.

And it can give you some AI powered debriefs so you can have like follow up conversations with the coach about your performance and get that live coaching.

So that feature already is pretty good. I'll probably

pretty good. I'll probably maybe we'll do another live event in a week or two, or maybe I'll just do a YouTube video. But that mode is very

video. But that mode is very close to ready.

You probably even saw it in my in my view here. I just need to make it look prettier. So and

then voice mode too.

So this one does use a lot more credits, but also opens a lot more possibilities. Like for

more possibilities. Like for example, if your learners are call center agents or something along those lines and they need to navigate software while talking to customers, we're going to be optimizing voice mode so that you can have these conversations while also navigating software and it should provide some really realistic practice opportunities.

So I want to see what y'all built. So please share your sim

built. So please share your sim tag me on LinkedIn. We are

looking for sims to showcase.

So if you're creating something that can be shared publicly again, please tag me. I love to showcase some of your simulations on my site.

I can't even imagine some of the creative stuff y'all can do.

I think I've seen some ideas already in the chat. So I

really am excited to see that.

Let me see y'all want to see my fun screen I built around this.

It's your turn. Frank's

intrigued. So there we go. And

I think that brings us to Q&A.

Yes, it does.

So thank you everybody for making it make it to the end here. I'll probably hang around.

here. I'll probably hang around.

You know, we'll see how it goes over the next 10 minutes. I'm

sure I could hang around a little longer than that if y'all would like me to.

But let's dive into it. I'm

going to stop showing my screen.

And how are we feeling? I haven't

been able to keep as close of an eye on the chat as I would have liked. But let's see.

have liked. But let's see.

All right. Any questions? Okay

credits credits credits. I'm

going to add a page that talks about how credits are used.

Basically whenever you use the AI features in the tool credits are used. So if you have a

are used. So if you have a super long, you know, super long information inside of the editor that controls your AI.

Every time someone has a conversation with your AI, it's going to put all of that into the context. It's going to use

the context. It's going to use more credits.

If someone's chatting back and forth with your simulation five times, it's going to use a lot less credits than if someone's chatting back and forth with it 20 or 30 times per simulation.

So it's hard to put an exact number on that. I will be adding an update very soon where you can see the credits used per conversation in your dashboard.

And for each of the all of the conversation history where you can see the recent completions, you'll be able to see the total amount of credits used the credits used for the conversation itself and the credits used for the evaluation.

So we're gonna give you some good insight so you can start learning for your Sims for your audience. Just how many credits

audience. Just how many credits you might need to roll it out.

When if credits start approach start running out, you will get an email so you can set an email alerts to say, hey, notify me when it goes below a certain amount.

We can also set up like auto top up and things like that if or as needed. And then for organizations, we'll have like we could have like organizational bundles of credits across all of their designers and users.

Rachel, you've never seen me talk this fast. Yeah, trying to really deliver on that time promise, I guess here.

This can definitely be used in an LMS. So we are going to make it so that while embedded inside of a storyline course, it can report the results up to the LMS. So it can report the score and the pass or fail. And then also these simulations can be exported individually as X API or score packages. So you don't need to put this inside of storyline.

That just makes it more engaging, right? Because you

engaging, right? Because you can have a character moving around and stuff. But yeah,

these you can build basically a whole like learning pathway with these simulations without ever touching storyline if you wanted to.

Yes, you can. You can customize the chat box. So you so full color customization transparency customization. I

transparency customization. I need to add a font uploader setting option.

But other than that, pretty much everything else should be customizable right now.

Thousands of users, Jim. Yeah,

this. Yeah, I'm designing this to be used for thousands of users for enterprise users.

We would probably separate. We

would probably give them a completely separate instances so that they're not sharing resources with, you know, with with other accounts and things like that.

So you could set up your own database, your own users, things along those lines. So if

you have concerns around scale, yeah, talk to me. Let me know.

We can solve those problems, right?

Those would be a good problems to have. And with our

to have. And with our architecture, we can currently handle tens of thousands of users. So if not hundreds of

users. So if not hundreds of thousands.

So we are already using enterprise great approach so we can handle a lot of people. I'm

not saying no hiccups might come up along the way. That's

why we're still in public beta.

But yeah, the core approach already works for a lot of people.

Tom for enterprise. Let's let's

talk about enterprise. I'm

looking for some design partners, roll out some some pilots. I'm already having some

pilots. I'm already having some enterprise conversations. So,

enterprise conversations. So, yeah, TBD on that enterprise piece. But right now I'm

piece. But right now I'm looking for some early, early people.

Build up, build up some case studies, right, and give an early enterprise users direct impact on the on the roadmap.

That's where we're at right now with that.

The free option is 5000 credits total right now. Yeah, if I do want to have some sort of option for people who are using it for their portfolio. The

idea is that if you're using this to roll out for learners on the job.

We want you to pay for it right those features are the paid features essentially, but I do want to make sure that portfolio users have have some good options here.

So I'm open to any feedback around that.

Yeah, it quad is being used to power this right now.

Or I should say anthropics API is what's powering this right now.

Would I be able to import specific scenarios and details into the simulation that I already used in a previous storyline. Of course, yeah,

storyline. Of course, yeah, that would be great. You could

dump in some content in that scenario description piece be like hey I need someone to practice using these skills in a realistic setting dump in the content generate the SIM.

And it should it should create something pretty impressive and realistic for you.

And yeah, Emily, let me know if that ever becomes a concern 5000 I think should be pretty generous for hiring managers reviewing your project. So

unless your portfolio project is going like super viral or something.

And you know you'd like hundreds of people using it. I

don't think you should worry too much about running out of those credits, but if that ever is the case for some of y'all if you're running low on credits and you have this on your portfolio.

At least in the short term right send me a message I want to I want to make it work for you.

How this will work with Excel accessibility Kelly I haven't done a deep dive into into that yet you know we have like focus states and things along those lines but I haven't done a full review on right consequences and requirements for that.

You can email me at Devlin at Devlin AI, if any of y'all have questions. If you have product

questions. If you have product feedback right while you're working in the tool there is a feedback button in the bottom right please use that especially during the beta.

I'm very responsive to that I check it almost obsessively.

And again I've already made some of my favorite updates based on feedback people share with that with that option.

Once you do subscribe once you are a member you can buy batches of credits.

Question about selling this is safe to a cyber security company. Ask your IT department

company. Ask your IT department so get your you know I'm happy to talk to IT departments review security questionnaires you know talk data data processing agreements things along those lines so yeah again we are in the process of removing more and more, you know removing any roadblocks that orgs might have to using this at their org. I'm

very responsive right now I want to make it work. So,

nice Oh Kelly nice okay send me send me an email yeah anyone who's interested in enterprise stuff might need some customization things along those lines send me an email as soon as possible I'm having those conversations this week and next week to try to find out who our first partners will be.

So yeah I'd like to get a good range so government client would be great to add to the to the mix.

Aaron using copilot as your AI tool IT have any, any issues using this platform slash data security.

They might have requests. Yeah,

they might have requests that aren't in place right now because again this is a beta.

But we are again tackling those requests, one at a time so and each org might require something a little different so yeah I'm exploring how to handle that quickly.

Judy nice I'm glad to see you here your team's and ah okay great. Thanks for making it

great. Thanks for making it here and again thanks for your feedback it's helpful.

This can be embedded into rise.

Yes, that is one of the options I haven't done as much like customize you know there isn't as much.

Is it you can't like update completion in rise right now, based on how people are doing in the simulation but if that is something people need right if we want to update rise scores or update rise progress things along those lines that should be relatively light lift for me yes bring your own, bring your own key great question Eric that will be an option for enterprise clients so if you have a custom LLM

gateway. If you want to bring

gateway. If you want to bring your own API key that's going to be available at the enterprise level.

Jennifer I can't see any reason why this would conflict with articulates AI features. The

way this embed works is it's, it's its own completely separate thing. It just talks

separate thing. It just talks to storyline through those variables.

So yeah, articulates features are quite different there.

So enterprise license what is the starting credit amount Sheldon we are not there yet.

Again, I'm in this stage where for enterprise clients I'm, I'm working looking for some early partners and we're going to build the enterprise program together. So, yeah, for those

together. So, yeah, for those these early people I'm going to, we're going to see what their needs are, we're going to run some pilots, we're going to, we're going to put together a solid enterprise plan but we're not there yet so I'm just having those conversations over the next couple of weeks.

Patrick, when you pay monthly what's the best way to position myself to hiring managers I want hiring managers to know I'm using this.

If you're looking for roles. I

think you could you know show screenshots of the tool, talk about updates you made you know so you generate a sim, test the sim, iterate on the sim, so say you know I generated it I tried using this approach I found that this, the character wasn't acting right in this way so I made these updates.

So highlight your iteration highlight the problem that you're solving with the simulation.

Share some screenshots from the tool to show the behind the scenes work that you're doing.

And of course embed the sim, you know somewhere front and center on your web page so that people can experience that I think, while you know a lot of hiring managers if they still haven't seen this tool I think it might be a wow moment.

You know for them to see this for the first time, so it might be a good opportunity if you're looking for roles. Yeah if

anyone needs to head out feel free to head out I appreciate you all being here for so long I'll probably wrap this up shortly just trying to catch up with the chat and still scrolling my way down.

Eric you asked about virtual desktop I use parsec it is a lifesaver, parsec that's how I'm controlling my desktop downstairs from my Mac upstairs.

So you'd think you probably need a decent home internet connection but as long as it's decent not even great.

There's like near zero lag it's the best thing it was made for gamers.

Kostin does it work in Romanian.

I think that should be an option so as long as Claude is able to work in those languages as long as anthropic models work in those languages we can work in those languages, so adding other again if you just fill out all the settings in Romanian.

Check out the same and it will probably respond and work in Romanian, but yeah use the feedback form I'm happy to add you know multilingual support that will be a very light lift with how this is working.

Thanks everyone who signed off.

So to buy credits you can click on that credits plus button in the top right, and that will be our core plan with the 35 credits per month and that will be a $39 per month for life for everyone who joins in our early beta program this month over the next three weeks.

Oh yes, Jane that would be great Jamer Jamie rise test court course. Yeah, let me know

court course. Yeah, let me know if you want to help me make this work really really well with rise let me know your specific needs in there.

Right now it embeds fine it will look great. It's just what do we want it to control and rise.

And there freelance IDs using that $39 subscription, I will want to work with freelancers at some point because the tricky thing with that is as a freelancer right if you're hosting this same for a client there the clients usage is going to use your credits.

So we'll likely need an option for freelancers where you can let the client, you know, put in their own billing information if they would like, and then the client is the one who gets billed for the credits not the freelancer.

I've already been thinking about that I do definitely want to make this work for freelancers.

So, yeah, if you if you're thinking about using this as a freelancer once you have some interested clients, I'll get on calls with you will figure it out together.

But yes, for speech to text as it been tested on rise been running into an issue with rise blocking the microphone.

That might be tricky Judy, I, I had, I stopped my voice testing to prepare for this live event over the past two weeks, I was working through that with storyline with storyline we obviously have some easier options.

But yeah rise I haven't tested yet. So that will definitely

yet. So that will definitely add that to my to do list to do list when I get back into the voice testing.

How long did it take me to create this. Oh, I start I

create this. Oh, I start I recorded a video I'm like I have this idea.

I want to start building this on March 12. And it's currently April 9.

The amount of hours I've spent on this though I'm probably I'm embarrassed to say I've spent a lot of nights up until one a.m.

Let's just say it's been a bit of an obsession.

So, yeah, less than a month from when I started but hundreds and hundreds of hours.

If that answers your question, education pricing Brenda I actually haven't thought about that I am very intrigued if you send me an email or submit that feedback for me letting me know your situation I'm sure we can.

I'm interested I think we could work something out there that's a great idea.

Thank you everyone appreciate all the support I still haven't caught up to the real time, where we are in the chat.

If you don't finish all the credits as it accumulate. Good

question there.

Subscription credits the 35 K credits a month those will accumulate up to three months.

So once you get those subscription credits, they don't expire until three months later, so you can accumulate you know over 100 K credits on the subscription.

And then when you purchase credits.

They are valid for 12 months.

So again once you're on the core plan you can purchase credit packs, those are valid for 12 months, just so we don't have like endless liabilities on our books.

Really good for case managers at the nonprofit I work at great yeah I'd love to hear about the use cases you're thinking of the session replay will be available after this.

Thank you all for attending the workshops.

Flying over all the racial it's flying over your high expectations good okay I'm glad I know we were talking about it a little yesterday I'm like, let me know what you think.

Okay got see some freelancers interested for voice version can we include an option to change the accent so I'm using 11 labs voices for this, this prototype.

So if 11 labs has a voice that works for you. That's likely

what the first version of voice will roll out with just because those do have the best voices so I'm not doing custom voice model development I'm working with the best that I can find.

Yeah, hi to thanks for helping it come to life to you. Yeah,

hi to you saw it from its inception right when this started as something at a tech academy workshop where I was teaching people how to build something like this with quad code and it was, you know, obviously more complex than this.

And so that's when I was like, what if I tried to build this, build this tool myself to make it as easy as humanly possible for instructional designers to create these simulations.

And so that's what led me there.

Johnny 300 learners more than 35 credits per month.

Maybe maybe not. I think that depends. That depends. But yeah

depends. That depends. But yeah

I love to get in touch and find a payment plan and stuff that would work out for you.

Yeah, yeah we this why we're still in the beta period so if we do need to bump those month those credits every month, I'm open to doing that right if this isn't enough for most users.

And then you know if there are re ups required or a higher tier plan with you know that needs heavier usage, but you know if those 300 learners are all every single month using multiple simulations and the simulations go 10 to 20 turns.

Maybe not. If it's a little bit more sporadic usage and it's you know, a few simulations. I

lean towards yes but I think we just need a little bit more data to get a good get inside into that.

Tanya I'm excited to see your portfolio piece.

I'm, it's very fun to create a piece of software like this when there's so much excitement about it and when it seems like it's going to actually help people so thank you for all the support when I see y'all building stuff with it. That's

what makes it more exciting.

That's why I'm like, gladly staying up till 1am every night and prepping for these workshops.

So thank you for for the excitement.

And cost and you have 11k plus learners. Yeah, we have we have

learners. Yeah, we have we have options for that that's where the bring your own key is going to come into play. So if you have thousands of learners that's enterprise level solution we're looking at so we your org likely has their own API keys their own agreements with these AI providers.

So we would not be charging you for credits at that point, you would be using your own orgs pricing.

So yeah at scale, you're not going to use the credits this is for smaller scale at scale you're going to want to use enterprise bring your own key.

It's going to be much more cost effective at scale.

I will say also if the costs or credit usage is a concern, we can use a faster model during a big portion of beta we had a model we were using haiku for five, which is a faster and cheaper model, but some of the, it would hallucinate a little bit more.

So I decided to just stick to sonnet which is what's powering these now just because the quality is so good. And it

doesn't really make mistakes and all of my testing so there is a trade off there where it could be cheaper and a little bit faster to get the responses but the quality is less consistent.

So, you know, for a tool that's built all around the quality of the simulations, I figured the cost was worth it but that is something you can configure.

All right, thank you everybody yes the recording will be available no worries if you missed it. A question as we

missed it. A question as we import the web object, still the tokens or credits will be consumed for each response from Frank, yes. So for each message

Frank, yes. So for each message that the user sends to Frank and each response we get back, it will use some credits.

So whenever we use the AI engine to power this stuff.

That's where the credit usage comes in.

The trial credits do not expire.

So your account doesn't expire the trial credits do not expire.

So, I might make less trial credit less credits available in the future to new people who sign up once we get through this beta period.

But yeah, you don't need to worry about your trial credits expiring.

All right, so I've caught up with the chat I'm not seeing any new questions come in, did I miss anything or are you all ready to call it a day and hopefully get building I want to see what y'all create.

You're welcome. Thank you all.

And yeah, as you please the feedback is love here so if you're using the tool and it doesn't do something you want you have any sort of question. Use that

feedback back button send me an email at devlin@devin.ai. This

is my baby right now, so I want to make it as good as possible so yeah if you can give me that feedback on how to do that, that is extremely helpful for me.

And yeah any early enter into a state of enterprise users.

Let's hop on a call. Let's see

how we can make it work. I'm

not. Yeah, I don't think price will be a concern right I want to make it work on a run some pilots get some case studies going, make sure this can actually work at scale for big orcs.

So, thank you everyone thanks for the congrats on the new baby David.

The software baby, and my email is devlin@devlin.ai.

is devlin@devlin.ai.

devlin. Oh, Jeffy Yes, where can you try the new SIM created by Sabrina.

I shared it in my most in my second most recent LinkedIn post let me just grab the URL for you I'll share it here for those of you all who are still live in the replay it will of course be available.

But yeah, it's a pretty fun one if you haven't checked it out already it's called brush strokes and lies. Oh Nathan you beat me to it.

There we go.

Enjoy everybody will add that to the to the doc as well as a sample at the bottom.

Alright, we'll have a great rest of your weekend weekend I can't wait to hear your feedback on your thoughts, and I will talk to you all soon.

Thanks for trying out the tool.

Bye bye everybody. Cheers.