Episode #63: Closing the Exponential Divide: Preparing Manufacturing People and Workflows for the AI Age
U.S. Manufacturing Today host Matt Horine interviews Nikki Barua, CEO and co-founder of Flipwork.ai, about the human side of AI transformation in manufacturing and why many initiatives stall. Barua argues companies often start by broadly deploying LLM licenses without clarity on job-specific value, structure, or new habits, creating an “exponential divide” as AI advances faster than people adapt. She emphasizes shifting from an industrial-age, hierarchical task model to fast human–machine learning loops, redesigning workflows with AI at the center, and implementing governance and guardrails for agentic systems. Barua discusses worker identity and self-worth concerns, urging leaders to clarify what AI takes from a role versus the higher-value judgment humans should retain. Flipwork uses a diagnostic (Flip Factor) and 90-day sprints to build real workflows and certify “agentic leaders,” and offers a free assessment at flipwork.ai.
Links
- Flipwork Website
- Flipwork on LinkedIn
- Nikki Barua on LinkedIn
- Veryable Is Revitalizing U.S. Manufacturing
- Sign Up on the Veryable Platform
Timestamps
- 00:00 AI Needs People First
- 01:51 Nikki Barua Origins
- 04:23 Why AI Adoption Stalls
- 07:33 Human Machine Learning Loops
- 10:24 People Squared Framework
- 12:11 Agentic Workflow Redesign
- 15:57 Identity Crisis at Work
- 20:03 Managing Hybrid Teams
- 23:18 Skills for AI Era Careers
- 25:59 Flipwork 90 Day Sprints
- 28:46 Assessment and Closing
- 29:17 Final Takeaways and Outro
Episode Transcript
Matt Horine: [00:00:00] Welcome to US Manufacturing Today, the podcast powered by Veryable, where we talk with the leaders, innovators, and change makers, shaping the future of American industry, along with providing regular updates on the state of manufacturing, the changing landscape policies and more.
Today, we're talking about something manufacturers can no longer afford to ignore, not the technology around artificial intelligence, but the human side of it. How do you get your people ready for a world where AI is changing the nature of work faster than most organizations can adapt? And what separates the manufacturers that thrive in this moment from the ones that will get left behind?
Our guest today is Nikki Barua, CEO and co-founder of Flipwork, a transformation partner helping organizations reinvent their people, culture, and workflows for the AI age. Nikki is a serial entrepreneur, keynote speaker, and best-selling author whose work has been featured in The Wall Street Journal, Fortune, Bloomberg, Forbes, and Fast Company.
Her best-selling book, Beyond Barriers, has been featured in Forbes, CNBC, and Bloomberg as well. At the heart of her work is a deceptively [00:01:00] simple idea, the most powerful technology on the planet is still the human being, and the organizations that win in the AI age will be the ones that figure out how to make their people exponentially more capable, not just their systems.
Today, we're going to talk about what the AI transformation really demands from manufacturing leaders, why most AI initiatives stall, and how to build an organization that can evolve as fast as the world is changing. Nikki, welcome to US Manufacturing Today.
Nikki Barua: Thanks for inviting me, Matt. I'm thrilled to be here.
Matt Horine: We are very excited to have you because this is a very hot topic right now. We've talked a little bit on our show about AI and had some good recommendations, but I think this topic centering around people is pretty unique because people, most of the time, talk about the latest update or technology. I know our team is constantly monitoring the next cloud update and those types of things, so this is a really great topic to get everyone's head around.
Just to provide a little bit of background, you've worked with a lot of Fortune five hundreds across industries. You had a pretty big career in, in working with transformation [00:02:00] across the board in those types of organizations. What was the moment that shifted your focus from transforming businesses through technology to transforming people, and how does your journey build that framework out?
Nikki Barua: I've had an incredible first row seat into watching how the world's most complex and large global organizations work their way through transformation, and typically the catalyst for any kind of transformation is some sort of technology, right? There's a technology disruption that happens, whether it's the Internet that led to e-commerce and all kinds of online capabilities and digital transformation to mobile and cloud, and every single one of those forced organizations to rebuild their capabilities for this new era Well, the through line that I saw was that while the technology kept changing and becoming faster or more scalable, the real unlock happened when those organizations were able to [00:03:00] empower their people.
Because in the absence of that, you're investing in the technology, and the people are getting left behind, and that becomes a one-sided capability where you may have the most sophisticated systems that have tons of intelligence, but you don't have the workforce that knows how to harness that, and you don't have the kind of culture that is operating at the speed of that technology.
And so for me, it became very clear that the part that needed a more innovative solution built for the age of AI needed something to really focus on harnessing the people power of organizations, because it's only when humans and machines learn how to work together and co-evolve together, that's when you become exponentially capable.
And that's really what led to the founding of Flipwork, and that's exactly the work that we do in the market.
Matt Horine: It's a really great perspective because I think what a lot of folks put this in a traditional framework, and AI is very different than changes that have maybe happened over the past thirty, forty years, and you take that [00:04:00] back all the way to the macro scale, look at the era of globalization, look at whatever, name the technology over the past twenty years, blockchain, IoT, whatever those things are.
AI is totally different, and people have approached it at the first it's a kind of a scary new technology, but now it's this augmentation that really makes a lot more sense, and the best organizations are thinking about it that way. I think you framed this really well. You've talked about and identified what you call the exponential divide or the gap between how fast AI is advancing and how slowly most people in our organizations are actually adapting to it.
We hear from manufacturers constantly that AI is a priority. I think its buzzword doesn't even do its service based on my experience, but the gap between intention and execution is pretty big. What's actually holding companies back right now?
Nikki Barua: The most common challenge that holds companies back is literally not knowing where to begin, and we've seen that over and over again with so many global corporations and [00:05:00] leaders where the mandate is very clear.
You have to become AI-enabled as a company. There's a lot of pressure from boards about investing in AI, and so everyone's feeling that pressure. The question is: Where do you begin? What's the right thing to do, and in what sequence do you do them? And the knee-jerk reaction tends to be, "Let's just get a whole bunch of ChatGPT licenses.
Let's deploy AI to everybody in the company." And then very quickly they find out that the adoption rate is pretty low And invested in thousands of licenses, and barely 30, 40% of the people are even using it. And the ones that do use it are using it like a faster search engine, just typing in to say, "Hey, what's the best place for me to go to dinner?"
That's not creating any business value for you if that's basically the use case for LLM licenses. And so that leads to then the next step, which is then they make the usage mandatory, telling everybody [00:06:00] in the company, "Hey, we invested in these licenses. Everybody needs to be using it." Or sometimes even creating incentives and leaderboards, getting people to use it more.
So then token costs go up because everybody's now just using it because they're being forced to What doesn't change in any of these things is really understanding what is the real value of AI in how I do my job. What do I need to do differently? How do I think differently? How do I create value differently?
People just don't know what they're supposed to do different. And so what ends up being the real bottleneck is that there's lack of clarity on what they're supposed to do different, and then there's lack of structure to help them shift from how they used to do things. And when that gap exists, that becomes an exponential divide because week after week, AI models are getting smarter and faster, and week after week, your people are still stuck in the old way of doing things, and they haven't changed any of their [00:07:00] habits or their mindset or the behaviors.
And so over time, that compounds to the point where the difference between the machine and the human becomes pretty exponential, and that is the real challenge for companies to close the gap.
Matt Horine: It's really well framed because I think we were talking about it and some of my personal experience when I've seen people have this shift maybe over the past 12 months, which is an epic in AI terms, right, is the idea that the initial 2022 version of ChatGPT was something like cheating on your homework or using it as a search engine.
But it's clearly become a lot more than that. You've... One thing that I did catch in your work, you said the industrial age mindset, which is moving tasks from inbox to outbox, which is what we typically think of work, accomplishing a set of tasks, is what AI actually breaks. How do you help manufacturing leaders understand what they, what this actually means for their operations and how it can enhance their operations?
Nikki Barua: Yeah. The real power of AI that needs to be harnessed is thinking about it as [00:08:00] powerful learning loops between the machine and the human, right? The faster you're able to do that, where the experienced professional is giving direction to the machine to accomplish a task, but then based on how the machine produces that output, it's able to refine, apply its judgment, use the years of hard-earned wisdom in refining that output.
So think of it even on a simple terms like when we are prompting Claude and asking it to do a task and it comes back with a response. If you simply accept the response as is, you're gonna get a pretty basic level output. But if you're refining that based on your experience and your wisdom and your judgment, then you're gonna keep iterating on that until it becomes a far more sophisticated output.
But that is the loop. That's the human machine loop of iterating over and over again until you get to something that is truly valuable, and that's the piece that is so different in how the industrial era [00:09:00] organization was designed like a pyramid. It was direction from the top and every layer just executed based on the direction they got and got to the next level of out- And what happens in that kind of pyramid structure is that communication flows slower, input to output takes a long time because it's going in a very hierarchical, linear way.
And when you take that structure and simply slap on the power of AI intelligence, you're adding horsepower to a machine that's totally outdated. You're taking a revolutionary tool and putting it in a completely outdated model. The opportunity that really exists for manufacturing leaders is to really think about how do you go from that kind of structure to a fast-moving org design To thinking of a giant army to a Navy SEAL team that has the same charter, that has the power of AI a-alongside it, but is moving fast, executing quicker, iterating multiple [00:10:00] times to get to that output.
When you're able to do things that way, that's where you shift into the intelligence era instead of the industrial era.
Matt Horine: That makes a lot of sense, mostly because those hierarchies are something that's very ingrained in everybody in most facets of life. Being able to discern and decide this is not something that's-- fits-- squares neatly into this one position.
That is a great way to frame it and think about it. Let's talk about the People Squared framework. I think that's a really important part of your work. You've built this framework, human creativity plus contextual wisdom multiplied by agentic scale, if I have that defined right. Those two things are the things manufacturers have been building on the shop floor in their own environments.
How do you help them recognize and preserve that value as AI changes the nature of it?
Nikki Barua: The reason this framework was created because one of the most common questions we kept hearing is, we understand that the real power of AI is about augmenting humans, but what does that actually mean? What does augmentation actually look like in [00:11:00] practice?
What parts do I let go, and what parts do I preserve? What this gets back to is there are things that only humans can do, being able to solve novel problems, being able to leverage our creativity and imagination and, and when you combine that with the lived experience, with the institutional knowledge that only you bring after thousands of reps of seeing problems, and now you're able to scale that with AI.
So if in the past you could only handle this amount of work or this much oversight of systems, what if you could ten-x that because you have let go of the things that AI can do the repeated decision-making or the repeated execution of something and elevate yourself to higher value work? When you're able to scale at that level, effectively you're getting a much higher output per person, per leader than you were getting before because you're releasing the capacity of things that machines can always do faster, better, [00:12:00] cheaper than a human can and redeploying that released capacity into higher value decision-making and judgment that you should never let go to a machine.
Matt Horine: People aren't ready to make that hand off. It's not something that does the work for them, and thinking about it from the enterprise perspective, agentic AI is becoming a real force in enterprise level operations, not just down to the individual And building systems that don't just answer questions, but they do take that autonomous action.
I heard you mention that a little earlier in the show. What does that mean specifically for manufacturing organizations, how that decision-making process becomes part of the loop?
Nikki Barua: Yeah. It all begins with looking at workflows. For anything that we do, there's a specific workflow, whether it's on the manufacturing floor or in related decision-making cycles and management work that you have to do.
Everything has a specific workflow. And so the opportunity really is about instead of simply looking at the current workflow and thinking about how to automate it, imagine if you had a [00:13:00] blank canvas and you could design an entirely new workflow with AI at the center of it. What would that look like?
What does that mean where you have a team of agents making those decisions, taking those actions, but have the governance and the guardrails provided by an experienced leader who knows, uh, what great looks like, who knows where risk exists, and who's able to override the decisions made by AI and make sure that it does not lead to unnamed risk?
And that's what ultimately it's about redesigning the workflows very thoughtfully. But the difference I've seen is instead of taking the current state and simply improving it, I would challenge by saying, "How might you do it if you invented it today for the first time? What would that look like if you design a workflow with AI at the center completely new, and then put all the pieces together?"
When you take that kind of approach, that's what agentic workflows are able [00:14:00] to unlock the kind of productivity and capacity that was never possible before.
Matt Horine: That's a really great way to phrase it because I think there's-- what's the saying? That there's nothing more dangerous than saying we're going to do things the way we've always done it, and rebuilding it with AI sounds like a horrible idea, right?
It's a really good chance to kinda take a step back and look at the process as a whole.
Nikki Barua: But some of the challenges that in, in practice really show up is that you can take an existing system or an existing workflow and make it super sophisticated. You can throw AI tools at it. You can automate decision-making.
You can create predictive systems. You can do all of those things. But then you have people that are using those systems or operating as part of those workflows. If you haven't retrained them on what outcomes are expected, what the outputs need to look like, when to trust the system, when to override it, when all of that capability is not built in, [00:15:00] you're essentially putting power into something autonomous with no guardrails in it.
And that's why the idea of designing it very thoughtfully from the ground up alongside the people that actually use these things and leveling up their capability and their decision-making skills, not just the s- power of the systems themselves.
Matt Horine: That does make a lot of sense, and it's a good transition to bring something up about AI initiatives and why they stall, which has been a pretty recurring theme.
I'm gonna cite a study that I saw without name or any data here, but it's flashed a couple of times. Somewhere around 90 to 95% of business leaders thought that their AI initiatives weren't either generating revenue or getting an ROI for what they'd wanted. I'll have to find it and link it in the comments at some point, but it st- but it stands out because that's 9 out of 10.
Most AI transformation efforts fail to deliver what they promised, and I think you've identified a couple of reasons why. But one of the most powerful things you've identified, and it's [00:16:00] pretty unique, I haven't had anyone on the show quite say this, is the identity crisis that happens when workers ask, "If the machine can do everything, who am I and what am I doing here?"
And so how do manufacturing leaders address that question effectively, which is the opposite of the initial adoption, right? What happens if you get too far into it and they don't know where their place is?
Nikki Barua: Yeah, that, that study came out, I think, sometime last year, and I, I believe it was an MIT study that talked about 95% of AI initiatives did not yield any ROI.
A lot of this was based on pilots and, like any new technology, the first reaction that companies have is, "Let's try this out. Let's roll out a pilot." And the, some of the reasons why those pilots did not yield ROI is that because it, the objective wasn't about business, the objective was almost like testing out and playing with the technology.
No matter what you're, what technology you're looking to adopt, the most important thing is connect it to a business objective. What do you really wanna [00:17:00] achieve with that? Do you want higher profits, greater productivity, more efficiencies, more customer e- experience? What does that look like? What's the clear objective that you're aiming for?
And then align the initiative to achieve that. We have a proprietary diagnostic called Flip Factor, and it measures the agentic readiness of the workforce and the organizations. And what our data has revealed is that even though there's a lot of sort of perception around people resisting AI, they're like, "Oh, people are afraid of being replaced by AI, so they're resisting AI."
What we found was that's not the case actually. They're not resisting AI. They're actually excited about AI. What they're resisting is the loss of self-worth when they feel like they're unclear about their own value and their identity in relation to AI. We all grew up in an environment where in order for us to be the best at our craft, we simply had to compete against [00:18:00] another human You could be in a classroom, and if you wanted to be number one, you just had to beat the other students, and then you get to work, and if you wanna get promoted, you just need to be better than your peer group.
We were always competing with humans, and in that, our identity was very clear of what we needed to be good at. The challenge people are facing right now is that it's unclear to them when it's me against AI, where am I good? Where do I excel, and how am I supposed to do the computation as fast as AI, or how am I supposed to analyze as fast as AI?
So you start to question, where does my value actually lie? And if they're not provided that clarity by their leadership, then it leads to the subconscious sort of stalling of saying, "I still wanna feel valuable. I still wanna feel like I matter." And that identity crisis, that is the real bottleneck that actually prevents people from becoming truly agentic.
And until leaders are able to sh- help [00:19:00] them see the path, to say, "Hey, AI is only gonna take the parts of your job, not your entire job, but the parts of your job that frankly almost looks like grunt work. Why do you wanna keep doing paperwork? Why do you wanna do data entry? Why do you wanna do basic analysis?
Why not let the machine do it? And with your freed-up capacity, we're gonna elevate you to doing the things that are your zone of genius. That is your superpower, and do more of that." But that needs to be explicitly stated by leaders because in the absence of that, people are making assumptions around the things that they don't understand how their identity shifts.
Matt Horine: That's pretty spot on for talking about these types of large scale transformations, and particularly thinking about it in the structure of the task doing that we were mentioning earlier in the show. You've talked about the shift from being excellent task doers in that kind of grunt work, which is just stated bluntly it is, to becoming outcome orchestrators, which is maybe taking a little liberty here, [00:20:00] people doing the best at what they do best, or the part of the job that they maybe enjoy.
What does that transition demand for manufacturing managers and supervisors specifically? Because sometimes that middle layer of management or that shop floor management, empowering that can be a little bit tricky.
Nikki Barua: Yeah. Pure people management, the way that used to exist before, I think we're gonna start to sh- see a shift in that.
I, I think because we should expect to have hybrid teams Of humans and AI agents. As a leader, you're managing, or as a manager, you're responsible for both humans and machines together, and your job is to ensure the right outcomes, not just an output from that. So in the past, if you're used to running a plant and managing P&L or leading a team through crisis, you're used to making those decisions by yourself.
But if you ask someone, how are you making those decisions with AI as your partner in that process? A lot of them [00:21:00] haven't made that shift. They don't know what that looks like. If you ask, how do you, how are you managing your team differently from just having a team of humans to having a team of humans plus agents?
Those all require us to do two things Number one, unlearn the old way. Because as long as you don't shed the old way of doing things and thinking about things, you're not gonna learn the new. But then the second challenge then becomes of, like, how do you create an entirely new structure and a playbook that elevates your capability as well, and that you're not a bottleneck in that process where you simply don't know how to do things differently, so you don't do it at all.
Matt Horine: Yeah, that's a totally different mindset, right? The idea of having that kind of a hybrid team is a lot to think about for folks who are just getting introduced to the concept of having AI in their everyday life and on the shop floor.
Nikki Barua: One thing to keep in mind, though, in that is that in many ways, just think of managing your team of AI agents no differently than managing a team of people.
[00:22:00] You're simply-- You still need the same kind of, like, job description. What do you want this agent to do? You still need to have performance expectations and benchmarks. Is-- When is it meeting expectations? When is it failing expectations? You still have to give it feedback, so it keeps getting better, and you still have to define where the guardrails and override capabilities are.
So it's no different from managing a team of humans. Where I've seen it go sideways is where leaders think, "Oh, it's AI, it's smart. I don't need to direct it. I don't need to oversee it. I can just get it going and set it up, and it's off on its own," or, "I don't need to think about the interaction between the people and the machine."
So all of those things, think of it just like you would design any other team. If you had a team of 20 people, you wouldn't just throw them on the plant floor and tell them to go figure it out. You would give them explicit direction. You would set the guardrails. You would tell-- You'd know the escalation path.
You would identify the risks. You would make sure that there's [00:23:00] collaboration amongst them, that the handoffs are right. So all of those things still apply, and it's really understanding that just because something has extreme intelligence doesn't mean it's gonna have the same kind of judgment and decision-making as you do, and that's why the role of leaders is even more important now.
Matt Horine: For sure. It's something that, you know, considering both ends of the workforce on the leadership side but also a workforce that is trending towards starting a career in manufacturing, it's a lot of what we hear about in the re- reshoring movement, the reindustrialization movement in the United States.
Manufacturing is getting a good look for as a career path for the first time in a long time. A lot of people have been doing it, but explicitly manufacturing for the purposes and the pursuits of reindustrialization is something we talk about a lot on the show. For a young person considering a career in manufacturing, what skills should they be developing to be irreplaceable in the AI age?
Nikki Barua: It's all the human skills that cannot be automated, simply put. But here's the [00:24:00] exciting part for anyone considering a career in manufacturing When you look back to previous generations, kids were raised to pursue those white collar jobs and professions where it was like, be a doctor and an engineer or a lawyer.
Those are the very professions that are most at risk because AI can do all of those things. And yet, s- I think it was a Deloitte report that said 80% of manufacturing tasks are still gonna be human-led. So there's tremendous value in what is not only a growth industry, because the future is gonna require a ton of manufacturing, but it's also gonna be something that positions you to bring unique value and not just get automated away.
So it's an exciting field to be in, m- but the part that will make you truly irreplaceable is honing those skills that are innately human. So whether it's relationship skills, whether it's creativity, it's being able to address a customer's problems or resolve [00:25:00] conflict, inspire people and lead teams, all of those kind of skills are the things that are never gonna go away And so doubling down on that and building that, what I think of as an E-shaped builder, right?
Like, you-- I think that's the key is you cannot be so narrow in your skill set right now that you just have depth in one area. You need to bring the experience. That you wanna keep building the experience. You wanna develop expertise in a specific area, and you wanna keep exploring with a curious mind at other areas that are emerging, so you never get stuck and complacent in just one line of work.
As long as you're operating against all of those areas, you're gonna have a high surface area for continuous growth and innovation. That's the only path to, like, having great careers in this age.
Matt Horine: Timeless advice, but very, very prescient for what's going on with AI. But it's just fundamentals sometimes, right?
People tend to overlook that and not value it as much as they should. Let's talk a little bit [00:26:00] about Flipwork and what you're doing there. You offer a 90-day transformation model in the age of AI and the speed of AI, right? Why 90 days, and what makes that timeline realistic for organizations that have been operating the same way for decades?
Nikki Barua: Look, as someone who's spent a career in transformation, I was used to multi-year transformation programs. They would have a start date and an end date, and it would take months or years to achieve an outcome. But if you apply that same model in today's day and age, by the time you finish the transformation program, that world has changed 10 times in the midst of it.
So you just can't afford that kinda slow timeline. I think companies have to go from looking at things with this one to three-year mindset to really looking at how do you unlock agility? Because being able to adapt to change is literally the greatest superpower you can have right now, and that means your entire organization, your workforce needs to become continuous learners.
And [00:27:00] so the approach that we take is, one, we use the diagnostic to measure every single individual's AI readiness and exactly where they are and where they need to be. So that gives a very clear picture of what does that look like for you as a whole company? Where are you in the spectrum? And then from there, they go into 90-day sprints where in 90 days they're not only mastering their AI skills and their competency, but they're actually building real workflows.
They're building real agentic systems. So this is not learning in theory, this is learning by doing. It's the kind of thing that you really get a deep feel for the power of AI augmentation only when you learn how to become bionic yourself. When you have that human machine collaboration, that's when something magical unlocks and people get it, and they're like, "Oh, now I finally get the power of what is possible."
And at the end of that sprint, they have a capstone project where they've actually rebuilt real [00:28:00] systems and real workflows And only when they meet that criteria do we certify them as agentic leaders. So that whole process, it does a few things. It helps people understand very quickly what is possible.
It helps them believe that they can change, that it's not this change happening, but they are able to change. And thirdly, it shows verifiable proof that they have made that change. And so when that kind of shift happens and you do that at an organizational level, it reinforces this mindset of continuous learning and continuous growth, and that's what makes them bulletproof navigating through this much uncertainty and change.
Matt Horine: It sounds like a great process. I think a couple things you said there was not only do they believe it, but they have the verifiable proof, right? And that's, I think, what most people are searching for in this age. If our listeners wanted to find out more about your work or how they could potentially do an assessment, where could they go?
Nikki Barua: Yeah. They should go to flipwork.ai, take the assessment. It's free. It gives you a personalized report that'll help you understand exactly [00:29:00] where you are and what you need to do next.
Matt Horine: Excellent. Nikki, thank you very much for joining us today. It's been a really insightful conversation, and I know a lot of manufacturing leaders are looking for more clarity in the age of AI and getting real results, which is what manufacturing's all about.
So thank you so much for joining us.
Nikki Barua: Thanks for having me, Matt.
Matt Horine: Manufacturing has always been about turning raw materials into finished goods, transforming inputs into outputs through skill, process, and technology. What the AI age demands is a different kind of transformation, and the manufacturers who figure out that equation will define the next decade.
Nikki and the team at Flipwork are on the front lines of that work, helping organizations close the exponential divide and build the human capability that no AI can replicate and no competitor can easily copy.
To stay ahead of the curve and to help plan your strategy, please check out our website at variableapps.com and under the resources section titled Trump 2.0, where you can see the framework around upcoming policies from the administration and how they will impact you as a manufacturer.
If you're on socials, give us a follow on LinkedIn, X, formerly Twitter, and Instagram. And if you're enjoying the podcast, please feel free to follow the show on Apple Podcasts, Spotify, or YouTube, and leave us a rating and don't forget to subscribe. Thank you again for joining us and learning more about how you can make your way.