Harvard's Raffaella Sadun on why it's so hard to become an AI-first organization

You know, we really want people to adapt to new technologies and that’s why you have so much excitement.

Reskilling is bloody hard.

And why is it hard?

Because we are asking our workforce to change who they are.

Welcome to WorkLab, the podcast from Microsoft.

I’m Molly Wood.

On WorkLab, we talk to experts about AI and the future of work, from how the technology can supercharge human capabilities to how it can transform entire organizations.

Our guest today on location is Raffaella Sadun.

Raffaella is the Charles E. Wilson

Professor of Business Administration at Harvard Business School, where we are today.

She co-chairs The Project on Managing the Future of Work and co-leads the Digital Reskilling Lab, where she studies how strategies can actually change day to day work and performance.

Raffaella is also a research leader on a new collaboration between Microsoft and Harvard Business School.

This is a pioneering effort that is dedicated to advancing research and practical learning on how AI can drive strategic business value and empower the workforce.

Raffaella, welcome to WorkLab.

Thank you for having me.

So, talk a little bit about your work, if you would.

You have been studying management and work since the early 2000s.

Can you reflect a little bit on what you have seen just in the past three years?

Yes.

So what we have seen over the past three years is in some ways very similar to what we saw in the with the past technological waves of ICTs and computers.

Meaning, there is this tremendous excitement and expectation that we are at the beginning of a new technological revolution.

And then at the same time, you also see the frustration I’m seeing.

Okay, but this major revolution, I don’t see it yet impacting productivity.

I haven’t seen it.

It’s everywhere.

You know, there is the famous solo paradox.

You see computers everywhere except for the productivity statistics, and it feels a little bit like that.

Now, what is different now?

I think what’s different is the promise of this technology.

AI has the capabilities of summarizing and allowing people to access tacit knowledge, that is typically, you know, exclusive to experts.

So now we have this incredible promise of making knowledge available and usable and potentially productive to everyone.

And this feels like a potentially revolutionary change for people and for companies, but it also entails some difficulties in how exactly are we going to introduce this incredible knowledge in what we do in organizations?

And that’s the I think that’s the challenge that AI poses to firms and workers.

It’s interesting that you say it has echoes of the past, and I wonder how I wonder how significant you’re finding it in terms of the work that you’re doing, the potential for gen AI, and then, of course, the reskilling that we’re going to spend a lot of time talking about.

I think the echoes from the past is that you really need to think about your own firm’s specific way of incorporating this new technology into what you do, how you create value, how your workers are organized, how people think.

And so, you know, in the past, we had this notion of the J curve of productivity, meaning when there is a new technology, you first see a reorganization, and maybe a dip in productivity and then things take off.

Here it feels like every firm is in its own J curve.

Some J curves look really deep with some amazing potential.

Others are, you know, starting to see what’s coming.

So it’s this incredible moment of trying and experimenting how AI could change what we do.

Right.

And we’re early.

We’re early enough to have an incredible impact.

We are incredibly early in this journey.

Yeah.

Okay, so we alluded to this partnership between Microsoft and Harvard Business School to develop what Microsoft is calling a Frontier Firm.

How do you think about a Frontier Firm and why is this such an important framework?

You know, I think of a Frontier Firm as a firm that first and foremost understands that this is the time to learn.

Now, the question is, how are we going to learn how technology benefits us?

And so for me, a Frontier Firm is a firm that is investing in the process for strategic experimentation and for taking the learnings from this experimentation back into the organization.

What are you seeing emerge?

So, as we try to identify firms that fit into this framework, that can be or will be or are Frontier Firms, what are the patterns that lead to, do you think, a successful reinvention?

You know, I think it starts from what are you about as an organization?

I think this is pretty much the vision and the direction that then helps to structure these investments in a productive way.

How are you going to be differentiated?

What is the value that you create?

I think that that’s, at a high level, having that clarity is incredibly helpful because what happens next is and then you start understanding that the experimentation is within specific strategic boundaries of value creation.

And it could be delivering a better product or delivering a better service, or doing things at a lower cost, but you need that type of guidance.

Different ideas are bubbling up everywhere, and I think that your role as an organization is to understand how do I make sense of all this activity in a way that comes back to me?

Then, there’s a second piece, which is, I think, very important.

Once you have understood what you want to do and how AI can create value for you, how are you going to bring the organization with you?

As we know, every time there is a major technological shift, there are some parts of the organization that will feel like winners and others where there will be tremendous fear of being replaced or even having to change.

And so I think that that’s like a second stage is building that buy-in and that adoption.

That is essential to make sure that the new ideas don’t just stay with the lab in the firm or with a few people who really see the future.

And then, let’s say down the line, we will get to the more managerial side of things, making sure that there is compliance and metrics and all of that.

But we are very far from that step now.

You know, it’s interesting.

This kind of trend of almost what you’re describing as personalization to a company really keeps coming up that that a technology exists, but if you try to apply it generally or you get distracted by every version and you don’t make it specific to your company, you can really go astray.

I see this.

I see this in the sense that it’s the part that is scary and at the same time exciting.

The scary part is there is no playbook.

I think we just have to accept that.

Right.

And there is a lot of desire for a playbook.

Just tell me what to do to implement this, I will win.

Let me... of course.

Of course!

I mean, when I create a new...

even for the most basic things, I think as humans, we need a playbook.

This is a moment of tremendous uncertainty, and oh my God, if only we had a playbook, that would be great.

The reality is we don’t.

And so I think that the issue is, how do we come up with something that really fits not only what we want to do, but our culture and our people?

And so, as you can see, this becomes from just being a unidimensional technological investment, this becomes something that is much more comprehensive.

It’s strategy, it’s people, it’s organization.

And this is where I see companies really being differentiated right now.

Is this part of the the work, do you think, that this partnership could accomplish between Microsoft and Harvard Business School or part of your research?

There may not be an individual playbook for every company, but certainly patterns will emerge, right?

I think that the the hope is that not only we develop understandings of what works and how things really work.

My hope is that it will also show the importance of collaboration within humans and machines, which is, you know, something that is very important not just for organizations, but for society.

I believe that what we might find out is that there is a playbook for learning, and we are used to talking about the scientific method.

There is no other moment in history where I think the scientific method, when it’s helpful because it helps you understand what hypothesis you want to test and understanding.

You know, this distinguishing between fluff and reality.

You brought up management and human and agent collaboration, and that is exactly where I want to go next, because that is a foundational element I think, of a of a Frontier Firm is the idea that every every employee, in a way, becomes a manager, now has an agent employee.

Yeah.

Talk about that mindset shift and how significant it is.

I think it’s huge.

But let’s start from the basics.

Even today, it’s not like everybody is a good manager.

I mean.

Well, just ask them.

They’ll tell you how good they are.

Yeah sure.

I mean, I’m sure that there is and often also the ability of being a good manager is not necessarily how successful you are from an individual standpoint, right?

I mean, I think good managers bring the best out of people, especially in places where there is team production, and that’s an asset, recognizing that there are qualities in managerial capabilities is a starting point.

Now imagine you’re adding an agent in the equation, which is potentially incredibly knowing and, has infinite memory, potentially, so it’s like this very eager, new person on board.

However, it doesn’t always get it right.

It’s very knowledgeable in some things and not in others, has a hard time telling you when it doesn’t know things, and needs specificity.

Right?

They need to know in it’s context.

And so I think this is an incredible resource that managers need how to manage.

Meaning, asking the right questions, understanding what can be delegated and what cannot be delegated, and also making sense of the answers.

So more than an employee, I think it’s more like having a very eager apprentice and having the time to onboard that apprentice and that apprentice learn how to behave.

I mean, I could really imagine everything that you’re describing also making for better management of humans.

A hundred percent.

Oh my God.

And we do need a lot of improvement there, too.

Yeah.

We’ll talk about middle management in a minute, when we talk about your paper and reskilling.

You have also, though, done a lot of research into how executives allocate their time.

Yeah.

What do executives and then I think on down the line need to change in order to optimize their use?

First of all, the same ways in which there are no playbooks on how to leverage this technology, I think that this is something that we, we have to have the humility to know different ways, different leadership styles of approaching the questions which are emerging.

My hunch is that it really depends on where the company is in its phases.

You know, at the very early stage where you’re trying to figure out the technology, I think a leader needs to be close to the action.

And that’s because you are potentially redesigning how you create value in a very fundamental way, and you don’t want to spend or allocate time and resources in things that are not necessarily what’s truly important for the organization.

Also, it gives you credibility, the ability to go down to where the action takes place is important at this stage.

Right.

There is a second phase, which is more of the adoption and the diffusion.

I think there it’s important to understand for the leader who is the key player to help you build buy-in in the organization.

What’s fascinating is that there’s so much difference in, especially in larger organizations, between the skeptics and the enthusiasts, not necessarily correlated with their technical expertise.

It’s more of a mindset, as we were discussing before.

So I think identifying who these people are, in that phase to build buy-in, it’s like a playbook of change management, and that’s pretty crucial.

To sort of restate it, find your internal heroes, your ambassadors.

Yeah absolutely.

And making sure that they have the credibility that is needed for for them to sort of bring the others with them.

I think it can potentially backfire to create metrics that are meant to capture compliance, when you are at a stage in which you are creating new processes and new ways of doing things.

You want to create metrics that are helpful to sustain this initial phase of exploration, which is not necessarily compliance.

Right.

Fascinating.

Let’s move on to reskilling.

We’ve so far been talking about this as a growth mindset and a learning opportunity, but reskilling is a very specific piece of this that lets people keep up.

A few years ago, you co-wrote this award-winning article for Harvard Business Review called Reskilling in the Age of AI.

I want to go into these a bit, one at a time, but you laid out five paradigms for reskilling employees.

Do you mind recapping those?

Totally.

I have them tattooed in my brain.

I’m sure you do.

But before we go there, I just want to say why, why we have these different paradigms, and I think that the fundamental starting point for us is reskilling is great, we really want people to adapt to new technologies, and that’s why you have so much excitement.

Reskilling is bloody hard.

I mean, that’s the other piece.

And why is it hard?

Because we are asking our workforce or people who sometimes don’t even have a job, to change who they are, to change what they do, change even their professional identity.

So if you state the problem in these terms, then you realize that reskilling is not people, is not putting people in a room and have them go through a PowerPoint or a passive training program.

But it’s something much more fundamental.

The first piece is you start by specifying why is it that you want your people to be reskilled?

What is the business rationale?

And can you help create some certainty for the organization that tells them, okay, this is why we’re doing it, I see it.

There might be a future for me because the organization is going in that direction.

I think it starts from that clarity.

The second piece is, I’m working through the hierarchy of the organization.

Imagine you have a CEO who’s convinced that the workforce needs to be reskilled.

And the second piece is this should not just be an HR problem, precisely because we are trying to achieve something that is bigger.

This is a shared responsibility of the whole organization, and especially at that very layer that reports to the CEO.

Now you go down one level and you see, as soon as things change, there will be some uncertainty, some parts of the organization will always resist.

And so I think here is where you think about, in particular, that layer, that middle management layer.

There are some incredible coaches and motivators in that layer, and there will be some other middle managers that see this as an anathema.

That’s something, you know, this is the additional thing I have to put on my plate.

So let’s try and understand how do we build that layer, that middle management layer, in a way that is cooperating and understands what we’re doing?

Most important within the firm is put yourself in the shoes of your employees.

You’re asking them to askew.

Is it clear enough for them?

This is a cost, right?

So is it clear enough for them what’s coming after they’ve learned, or they’ve made an occupational switch.

What’s the promise?

Because, right now, people only see the cost.

What’s the promise in terms of career?

What’s the promise in terms of incentives?

What’s the promise in terms of direction, which may be inside the firm or outside of firm?

A lot of this reskilling discussion now is done in a vacuum for employees.

Within industries, there is a lot of knowledge.

I think that the last piece is how do you put together the knowledge that is coming within your supply chain or within your ecosystem, so that you can learn and potentially cooperate on how to reskill large at scale.

Right.

Create a playbook.

Please!

Please yes!

I think you called that section “It Takes a Village.” Oh, yeah.

Oh yeah totally.

In of many, many industries, one of the really important pieces of this, as you’re saying, it’s the responsibility of every leader and every manager, and that gets to that kind of middle manager layer.

How do you build that sense of ownership across an organization?

I think you make it...

you make sure that the mentoring and coaching of people becomes part of the role of middle managers.

You know, they do it.

They do it every day.

In some places, this type of activity is not rewarded.

In other places, actually, it’s an expectation that middle managers are evaluated for how they contribute not only to the financial performance, but also for the development of their people.

And so I think at that level, it’s what do you reward?

And do you make sure that the creation of human capital in your firm, because that’s what we are doing, is something that is seen as currency in the organization.

And then talk about techniques that work and don’t work.

It was interesting to read that, for example, no one, All due respect to every corporation everywhere no one loves to sit in front of a computer and click through a training, that there are ways to accomplish reskilling that have better outcomes but also engage employees.

Yeah.

I mean, look, we have some studies that have done this also with population of workers that are being reskilled into the workforce, not just within an organization.

Imagine how complicated that is.

And these studies tell us that it’s typically much better to think about reskilling, not just on the technical side, but also there is a combination of hard and soft skills that are part of the program, so that’s the design of the program.

It’s the cohort.

Are you learning new things on your own, or are you part of a community or part of a movement?

And then lastly, I think it’s also understanding how do you feed reskilling into a journey, not just an episode that happens because that person in that office has bought this incredible software, and now I have to spend two hours clicking through the program, but it’s because we are trying to do something that is really important to the firm.

Yeah.

And in fact, point one in your article is that firms who are engaging in this reskilling need to see it as a strategic imperative.

It seems to me that what we’re talking about is sort of a bottom up reason for doing this, and one of the other points this piece makes is that you cited a BCG report saying that companies now are investing or could be or should be investing 1.5% of their total budgets, minimum.

They already are.

They already are. Yeah.

And so then if you think of it as a strategic imperative, does that give you the reason to keep increasing that investment?

That’s why it’s such an interesting topic to study.

The resources are there.

I think that companies are investing in training opportunities, and especially since AI has arrived, this push for upskilling and reskilling is there.

Now, the other piece, however, is are you doing it well?

And “Are you doing it well” is something that, I think every company should be asking.

I’ve seen this incredible training platforms. They look like Netflix, but they all start from the premise that people are happy to learn on their own and that they’re motivated to learn on their own.

That’s not true.

We are adults, and adults are not necessarily interested in learning new things or admitting that they need to learn new things.

And so I think that the opportunity here is to think about that amount of money, that there’s incredible resources, spend them on things that are really going to the to the core of the issue, helping people learn and practice new skills.

One way to do that is also by structuring.

We go back to the point of experimentation.

Run these investments or implement these investments with a scientific mindset.

Try to understand what works and what doesn’t work.

It’s not just enough to do a pilot.

You want to have hypothesis.

You want to have measurement and you want to learn.

And also design a program that is thoughtful.

A hundred percent.

Design a program that is human.

Right.

I think Right.

I am cognizant of the fact that we are essentially asking organizations in order, in the process of becoming a Frontier Firm, you will have to reinvent everything.

You will have to rethink your business imperatives.

You will have to reskill and you will have to reinvent your processes, which is hard.

Yeah.

And so I wonder how you think about organizations getting from pilot phase, from experimenting, to actually having sustainable process change that drives meaningful results?

Yeah, I think it starts from the recognition that it is much more than just buying a new technology.

And so, and getting into that mindset is not easy because it requires a big amount, a certain amount of humility to recognize that there is an opportunity for a change.

A little bit of paranoia, too, because you’re changing because you realize that if you don’t, somebody else will do it for you.

I think this is where it has to start with a very clear leadership mandate and understanding of why we are doing it and sort of direction.

That’s where I would start.

I think the second piece is also learning right?

Who do you learn from and how do you learn?

We’ve talked a lot about the processes that a CEO or a company can set up to learn internally, but the reality is that there is a lot of learning that is happening everywhere.

And so my advice would also be to really have a wide lens to see how organizations are adopting.

What are they doing?

And go to the go to where things happen.

Don’t just rely on the reports because that’s a little bit of what happened during the Industrial Revolution, if you if you think about it, was this constant exchange of ideas between science and practice.

And I think we are at that stage.

Interesting.

So don’t try to implement change in a silo in your company.

Never comparing notes with your peers, for example, is that what you’re saying?

Yeah exactly.

So this is where I would expect people to learn from their networks.

Peer learning to be incredibly important.

Trying things, understanding what works, but also what doesn’t work.

Making sure that this is knowledge that gets actually highlighted and leveraged.

I think sometimes employees, like you said, the middle management layer can be wary of experimentation that then doesn’t turn into something meaningful or where you’re playing a lot, but you lose coherence, maybe, from a business perspective.

And I just would love to know what you’re thinking about, how organizations should approach this implementation, this pilot process, or not, this bottom up groundswell.

At some point there can be also too much experimentation.

So I would think I would say within one important axis of your strategy or maybe 2 or 3 think about experimenting within these boundaries.

What I call the strategic boundaries.

I think it starts there.

The second piece is who is conducting the experimentation and making sure that you have people who are not just familiar with the technology, but also have an understanding of how the technology would get used in the organization or by by the consumer.

I think this is crucial because the technology cycle is very fast, right?

And people who are close to the technology are so excited, and they tend not to see all the frictions that might prevent that beautiful invention from actually being used by people.

So, getting at that level of interaction between technology and humans, but also the ability to quickly iterate and understand what’s working and what’s not working, that’s a capability.

Once you have that, then you can think of embedding the results of the experiments in the firm.

I have seen companies where the experiments were done so brilliantly, but so divided from the rest of the organization that when it came to the adoption and scaling phase, there was almost a dejection, as if it was a, a foreign object that you were trying to to put in.

So you have to build that third stage of adoption already in the early stages.

You want to make sure that that that it’s not perceived as something completely different from what we do.

I have spoken with other people on the show who have talked about that.

The danger of really siloing things in pilots.

Yeah.

You’re not doing that kind of grafting process that you’re talking about.

It’s like game theory.

Yeah.

You know, it’s really like game theory.

You know, if you put yourself at the end of the process.

That’s why I insist on declarative.

Imagine that you want to have massive adoption, and I’m assuming you’re not going to want the whole organization to turn.

You will want your people to adopt and embrace this, right?

So what would you need to have in place for that stage to go well?

What type of input do you, do they need to have?

What type of learning?

What type of checkpoints do you need to have to get there?

You seem to also be describing the process of sort of reverse engineering your metrics, your measurement.

Yeah yeah yeah I think that’s that’s part of the discovery phase.

But I think it’s reverse engineering, not just the technology, but how people and organization evolve.

Let’s do a little bit of tactics, a little bit of actionable Go ahead.

advice here I am a CEO.

I am absolutely convinced.

I’m sold.

I need to become a Frontier Firm.

I understand, The survival of the organization rests on this.

What are the three steps that you would recommend for me to get started?

Okay.

So I would say strategy session first thing, not months.

Like, you know.

Right.

Three things we want to do.

Okay.

Set your experimentation team.

Choose these people very, very well so that we have the technical, the human, the potential resistance and then communication.

Think about communicating what you’re doing, what you are trying to learn, what this means for the organization very well.

Is that technical enough?

That is so specific, it’s amazing!

for an academic?

Just pause it, back it up, write it down.

It’s just a very generic, because I think at the end of the day, you are going to have to tailor all of these things into, into your reality.

But that’s its own important takeaway, right?

Don’t try to do it the way that some other company did it.

Use these steps to determine who you are as a company and what will make you succeed or fail.

You should know, I teach strategy.

So...

Not only do I know, but I can tell.

Okay good.

Okay, so now I’m gonna ask you for one more piece of strategy.

I’m an employee.

Oh yeah.

And I want to make an impact at a Frontier Firm.

I also am all in.

What should I do?

I’m assuming you will need a little bit of technical knowledge there, because you need that in order to create.

The first thing is recognize that you are an inventor at that point.

If that’s what you want to do, you are an inventor.

So what does an inventor need?

I think you need some technical skills.

I think you need the same amount of clarity to understand how what you’re doing fits with what the organization is doing, and if not, these are good questions to ask yourself, right?

Is this a good fit?

And then I think you want to build a movement.

You want to be part of a coalition because it’s extremely hard right now in places where you have enthusiasts, and there are many companies where there are many enthusiastic people about AI, everybody’s doing their own thing.

And so there is revelation of ideas and my my feeling is that these ideas never get compared or aggregated in a way that we can really learn from each other.

So I think you are an innovator.

You have to have the technical chops.

You have to have a community where you can continue learning and and potentially your learnings will be useful to the organization.

And then finally, based on what you have seen before, what you’re seeing now, what you’re examining today, what do you think work starts to look like in 3 to 5 years?

How significant are these changes?

I think it’s very hard to answer the question, because one thing I’m certain about is that we are going to see massive divergence across firms. In the same occupation at firm X will be very different from the same occupation at firm Y.

And this is, if you like, consequence of these firms specific investments being made and people learning, tasks changing, and I’m expecting that there will be a lot of, at least in the very in the short, medium run, there might be divergence in how, how jobs are done across firms. That is fascinating.

I don’t want to open an entire new box of research, but what does that start to mean for employees who reskill.

At one firm?

I think that the match with the company is essential.

Yeah.

because this means potentially that you come in an organization and three years from now, your job will be different.

This is not for everyone.

That’s why the match between who you are and what your aspirations are, and where you work is fundamental at this point, because if you are part of a firm, you will, that is changing constantly or that is about to embark in this path, your job might look very differently.

That seems very empowering in a way for, you know, we talk a lot, we obviously speak to business leaders on this show and decision makers, but it is that’s an interesting way to think about it from the employee level, but also the hiring level.

Yeah, I mean, at the hiring level, this means that you are not hiring for the skills that people have today, but you’re hiring people for their ability to learn new skills, which is something that I think businesspeople always say they want and they appreciate, but often they conflate it with experience.

And this is not necessarily the same thing.

I’m going to stop us there because I have to, but thank you again to Raffaella Sadun, the Charles E. Wilson

Professor of Business Administration at Harvard Business School.

I so appreciate the time.

Thank you very much.

Thank you so much for listening to WorkLab and joining us at the Harvard campus today.

We’re so grateful to them for having us.

You can learn more about the podcast and the Frontier Firm AI Initiative between Microsoft and Harvard at Microsoft dot com slash WorkLab.

If you have a question or a comment about this or any episode, please email us WorkLab at microsoft dot com and check out Microsoft’s work trend indexes and the WorkLab digital publication, where you’ll find all of our episodes, along with thoughtful stories that explore how business leaders are thriving in today’s AI era.

You’ll find all of that once again at Microsoft dot com slash WorkLab.

And as for this podcast, of course we want you to rate, us reviews, and follow us wherever you listen.

It helps us out a lot.

The Work Lab podcast is a place for experts to share their insights and opinions.

As students of the future of work, Microsoft values input from a diverse set of voices.

That said, the opinions and findings of our guests are their own and they may not necessarily reflect Microsoft’s own research or positions.

WorkLab is produced by Microsoft with Trifilm, Sharon Callender and Matthew Duncan produce the podcast with editorial support from John Eilenberg.

Jessica Volker is the WorkLab editor, and I’m your host, Molly Wood.