Category Archives: Leadership

Motivating the Unmotivated

Motivating the Unmotivated

GUEST POST from David Burkus

Motivation can vary wildly on a team. At any given time, a few people might be highly motivated, while others are totally unmotivated. Ideally, there are times where everyone is motivated at once, but sadly there may be times when everyone is demotivated or burnt out. All this means that an inescapable part of a leader’s job is to motivate the unmotivated.

The good news is that leaders don’t have to rely on raw charisma or the inspirational words of a halftime speech from insert-your-favorite-sports-movie-here. Instead, motivation is less about the qualities of the leader and more about understanding the needs of the team and of each individual on the team.

In this article, we’ll outline five ways to motivate the unmotivated.

Change Up Tasks

The first way to motivate the unmotivated is to change up tasks. Novelty can be a powerful motivator, and the lack of novelty in a job can be demotivating. Few people get excited about coming to work and repeating the same few tasks over and over again. People want new experiences and new challenges. They want to feel that they’re making progress and they often judge that progress based on the projects they’re being given and whether those projects require them to learn new skills or merely execute the same routine functions.

As a leader, this means examining the task list of your motivated team members. Are they doing the same old over again or are they being given new, growth-inducing tasks and projects to work on. You may not be able to change their job description, but you can help them find new learning opportunities or ask them to sit in on meetings they’re not regularly a part of. Even a little novelty can go a long way toward restoring motivation.

Build New Bonds

The second way to motivate the unmotivated is to build new bonds. Over four decades of research have made a compelling case that relatedness is an essential element of intrinsic motivation. People want to feel cared for and feel that their work cares for others. They want to feel connected to the people their work serves and the people they work alongside. And if they feel disconnected or isolated from the team or the customers/stakeholders of an organization, they can become unmotivated.

As a leader, there are two ways to utilize relatedness to motivate the unmotivated. The first is to make sure team members feel connected to each other, most often by making time for socialization and connection through non-work discussions. (The second we’ll cover in a moment.) It may seem like a waste of time, but social functions, icebreakers, or any other activities where people talk about their lives outside of work create opportunities for stronger connections to form. And there’s a strong connection between social connection and motivation.

Re-frame The Work

The third way to motivate the unmotivated is to re-frame the work. As discussed above, knowing how your work serves others can be a powerful motivator. But for many jobs, teams are so far removed from the end customers or even from other teams who benefit from their work that they lose sight of how their work makes a difference. Their work loses task significance, and their motivation quickly follows. And the larger the organization, the harder it is to keep task significance.

As a leader, restoring task significance and relatedness requires re-framing the work or rebuilding connections to those who benefit directly from your team’s work. This could be by bringing customers in to meet your team, or by sharing thank you notes or stories of how the team’s tasks enabled others to work or live better. The test for whether your team needs a re-frame is how quickly they can answer the question “Who is served by the work that we do?” And if they can’t find an answer fast, they likely can’t find their motivation either.

Provide More Feedback

The fourth way to motivate the unmotivated is to provide more feedback. Ken Blanchard was right, “feedback is the breakfast of champions.” We already covered how a feeling of growth and development contributes to motivation. But without regular feedback, your people don’t know what to improve upon—or if they’re improving. As well-intentioned as annual reviews are, they are not a sufficient source of feedback to keep people motivated to improve. Instead, try regular check-ins and feedback sessions both individually and as a team.

As a leader, it’s important to note the distinction between providing sufficient feedback and becoming a micromanager. As people grow and develop in their role, feedback should shift from telling people how to do specific tasks and towards coaching them to solve problems they’re already equipped to solve. In the beginning, provide feedback to help them grow. But as they develop, provide feedback that helps them notice their growth.

Watch The Stress

The fifth way to motivate the unmotivated is to watch the stress. Most leaders know that too much stress can demotivate anyone. But too little stress can be demotivating as well. Psychologist have long known about a concept called eustress—the sweet spot of stress where the demands of the moment match their ability and capacity. Too much demand leads to distress and burnout, but too little demand leads to boredom and…burnout.

As a leader, watching the stress means monitoring your team’s capacity so that they don’t get overloaded. In which case, you’ll want to find ways to offload certain projects or otherwise reduce the workload. But it also means watching each individual on your team for signs that they’re not being challenged enough. In which case, you’ll want to consider other methods in this article for ways to help them feel more growth and challenge in their work. In either case, the goal is to continue to make adjustments and continue to watch the stress, to bring it back to that eustress level.

And monitoring and making adjustments is really the ideal for each of these methods to motivate the unmotivated. Because motivation is individual. It’s felt on an individual level. Which means increasing motivation requires knowing each person individually and continuing to monitor their motivation levels for individual adjustments that need to be made. But when you do, it will raise the overall motivation on your team, and raise the level of performance until everyone on the team can do their best work ever.

Image credit: 1,300+ free quotes for your presentations at http://misterinnovation.com

Originally published at https://davidburkus.com on January 2, 2023.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Thinking From No to Yes for Top Line Growth

Top line growth strategies and product applicability frameworks

GUEST POST from Mike Shipulski

Bottom line growth is good, but top line growth is better. But if you want to grow the bottom line, ignore labor costs and reduce material costs. Labor cost is only 5-10% of product cost. Stop chasing it, and, instead, teach your design community to simplify the product so it uses fewer parts and design out the highest cost elements.

Where the factory creates bottom line growth, top line growth is generated in the market/customer domain. The best way I know to grow the top line is to broaden the applicability of your products and services. But, before you can broaden applicability, you’ve got to define applicability as it is. Define the limits of what your product can do – how much it can lift, how fast it can run a calculation and where it can be used. And for your service, define who can use it, where it can be used and what elements without customer involvement. And with the limits defined, you know where top line growth won’t come from.

Radical top line growth comes only when your products and services can be used in new applications. Sure, you can train your sales force to sell more of what you already have, but that runs out of gas soon enough. But, real top line growth comes when your services serve new customers in new ways. By definition, if you’re not trying to make your product work in new ways, you’re not going to achieve meaningful top line growth. And by definition, if you’re not creating new functionality for your services, you might as well be focusing on bottom line growth.

If your product couldn’t do it and now it can, you’re doing it right. If your service couldn’t be used by people that speak Chinese and now it can, you’re on your way. If your product couldn’t be used in applications without electricity and now it can, you’re on to something. If your service couldn’t run on a smartphone and now it can, well, you get the idea.

For the acid test, think no-to-yes.

If your product can’t work in application A, you can’t sell it to people who do that work. If your service can’t be used by visually impaired people, you’re not delivering value to them and they won’t buy it. Turning can’t into can is a big deal. But you’ve got to define can’t before you can turn it into can. If you want top line growth, take the time to define the limits of applicability.

No-to-yes is powerful because it creates clarity. It’s easy to know when a project will create no-to-yes functionality and when it won’t. And that makes it easy to stop projects that don’t deliver no-to-yes value and start projects that do.

No-to-yes is the key element of a compete-with-no-one approach to business.

Image credits: Pixabay

Subscribe to Human-Centered Change & Innovation WeeklySign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Is it Possible to be Incorruptible?

Is it Possible to be Incorruptible?

Exclusive Interview with Eric Ries

This candid, wide-ranging Q&A dives deep into what Eric Ries calls the “physics of organizations” — the hidden structural and financial forces that dictate whether a company thrives or decays over time. Moving past superficial business trends, the conversation tackles the intense psychological toll of entrepreneurship, the systemic flaws of shareholder primacy, and the historical reality of alternative corporate governance.

Over the last two decades, Eric Ries’ ideas about continuous innovation, long-term thinking, governance, and market reform have reshaped company building and management practices. He is the creator of the Lean Startup method, and the author of the New York Times bestseller The Lean Startup; The Leader’s Guide; and The Startup Way.

Eric RiesAs a founder, he has put his own ideas into practice with The Long-Term Stock Exchange (LTSE); Answer.AI, an AI R&D lab; Virgil, a legal services startup; and IMVU. On The Eric Ries Show, he talks with world-class technologists, thought leaders, and executives building for the long-term. He lives in the San Francisco Bay Area with his wife and three children. He is excited to announce his latest book Incorruptible: Why Good Companies Go Bad… and How Great Companies Stay Great.

Ries offers a provocative look at how truly resilient, mission-driven institutions can protect themselves from the gravitational pull of short-term financial systems to prioritize long-term human flourishing.

Below is the text of my interview with Eric and a preview of the kinds of insights you’ll find in Incorruptible presented in a Q&A format:

1. Why do purpose-driven companies create so much value for society?

The evidence shows that purpose driven companies outperform conventional companies financially as well as in almost any other dimension you care to measure, including the social dimension. Intuitively, this makes a lot of sense, because entrepreneurship is very difficult. Everyone says they know this, but I don’t think we really grapple with this fact nearly enough. If you just want to make money, there simply are better, more convenient ways than entrepreneurship. So to get not just the founder, but the early team, the early investors, all these people to take a risk to do this crazy thing generally requires some kind of extra-financial purpose or goal. Sometimes we call that vision, sometimes we call that, in a more demeaning way, strategy. But it’s also fine to call it purpose, which is really what intuitively makes the most sense to people that do this. This is one of those cases where intuition and the evidence agree, yet it is somehow still considered a controversial fact.

2. What were some of the most important lessons you absorbed during your time on the bathroom floor?

As I’ve been going around talking about the book, this is one of the stories that actually gets a very different reaction depending on whether I’m talking to an entrepreneur or somebody else. Entrepreneurs all recognize this moment, where I really thought my company was going to fail and I couldn’t handle it. A lot of non-entrepreneurs don’t get it. They’re like, “Why? It seems like a bit of an overreaction. Okay, you had a business setback. We’ve all had career setbacks — what’s the big deal?” But what you don’t realize until you’re in it is how much, especially if you’re doing something out of a sense of purpose or passion that’s personally meaningful to you, you start to identify with it and start to become inseparable from it. So that story is very important in the book because I learned a lot of important business lessons. I thought the company was going to die, but it didn’t. It survived precisely because of its mission, not in spite of it. I learned, in a very visceral way, about the forces, that prevent reform from coming to fruition in so many areas of our life, not just financial. And of course I learned a personal lesson about the importance of equanimity and the need to tackle the psychological and even spiritual dimensions of entrepreneurship if we’re going to create real change in the world.

3. How much chance is there of us getting companies to more broadly redefine profit to include elements of maximizing human flourishing?

This question reminds me of a of an incredible video of the great Steve Jobs before he died. He’s being interviewed at an industry conference at the time of the launch of the iPhone, when the Blackberry was the dominant smartphone in the world. It had something like 80 or 90% market share. A journalist asks this question something like, “Do you really think realistically you can take share from this dominant player?” And you can tell Steve is irked by this question, and I’m expecting because we all know his famous temper, that he’s going to lash out at the person. But he doesn’t. Instead, he says, “You know, that’s not really up to me. My job, our job at Apple, is to make the best phone we can, the one that we’re proud of. Market share is up to the customer. That’s their decision, their choice. We don’t think about that, we don’t know, and we don’t need to know in order to do our best work.” I’m paraphrasing because I haven’t seen this video in a long time, but that’s how I feel about this, too. I get this question a lot because people want to feel like, if I’m going to jump on the bandwagon, I want to know that it’s going to work. But the truth is none of us know what’s going to work, even those of us who advocate for these ideas. You, who’s reading this, are the only one who gets to decide if this is likely or unlikely. This is not what the economist John Maynard Keyes called a beauty contest. You don’t have to worry about what everyone else is going to do. You only have to decide for yourself if you think this makes sense to you. And if it does, well, like I said — like Steve said — it’s up to you.

4. As America becomes more capitalist and less of a free market economy, what steps can we take to reverse the regulatory capture, lawfare and other methods that degrade competition, purchasing power, class mobility and the American dream? Do we need a pCombinator? (purpose-driven company accelerator)

You’re asking questions about words that we no longer have consensus about what they mean. What is a free market economy? What is capitalism? What is regulatory capture? The very definition of these words is what’s under threat. If you look at the broader media landscape, the political landscape, in many, many pockets of our society now the very idea of a for-profit company is being attacked as inherently exploitative or extractive. The consensus that we used to have that we can be working commercially to improve the world and make it a better place, that used to be seen as quite obvious and now that whole idea is under threat. I don’t blame the people doing the attacking, especially the young people who have, after all, lived their whole lives, under this regime of a very extractive flavor of capitalism that goes by the anodyne-sounding name “shareholder primacy”. This is the simple idea that customers, employees, communities all exist as resources to be mined for the benefit of shareholders. But this question is also loaded with so many other political issues of our time that we are going to have to tackle if we’re going to come out of this darkness, as our grandparents who battled fascism once had to do. So, I don’t think it’s going to be as simple as fixing one thing. But I think that one of the things we have to do, among many, is build a power base, an economic gravity pulling towards the values aligned with human flourishing. And many of the political, economic, and social challenges of our time are downstream of this action in the same way that the catastrophes that we’re currently living through are downstream of what seem like very simple and relatively benign policy changes from the past century.

5. What should purpose-driven companies look for in a CEO as the company outgrows or outlives the founder(s)?

IncorruptibleThis is a really important part of the architecture of institutional longevity. Most companies fail the test of succession. The evidence seems to suggest that people who train and hire from within have a big advantage here. I think that is something we don’t even really teach anymore as a corporate value, but that is actually super valuable. There’s a reason why that old story of the employee that worked their way up from the mail room was such an important legend in the previous century. Now we hardly tell stories like that anymore. We tend to want the big fancy turnaround, the bold new strategy, the external CEO, which for companies that are in crisis makes sense. And since our modern best practices tend to ruin companies, they tend to be in crisis quite a lot. But what we want to do is we want to find a CEO who combines two really important elements. One, they personally, deeply and profoundly reflect the ethos of the company. This is why a company that doesn’t have an ethos can never pass this test because they don’t even know who to pick. But you don’t want someone who, who apes the values of the past, or is slavishly loyal to the specific things that worked in the past. You need someone who is both deeply aligned to the ethos, and who nonetheless is very performance oriented, meaning they see that when the ethos is working, it should generate long-term performance. They can’t get distracted by short-term blips but they have to have the adaptability to realize when sacred cows need to be challenged. Now, it’s commonly said that only a founder can have the moral authority to do this unique combination of things I’m describing, only they can go into founder mode, as it’s called. But I don’t think that is supported by the evidence. When companies have the right structure, they actually can imbue subsequent generations of managers with this moral authority.

6. Why is magnetic alignment so important for purpose-driven organizations and their survival?

I conceived of this book as a look into the physical forces, the underlying forces, that affect organizations. So not the surface level characteristics that we spill so much ink about, org chart, culture, business model strategy, even vision, things we can touch and taste and control. Those things are important, don’t get me wrong. But there is a deeper layer to this, like a physics of organizations. In the book, I explore very dominant force that I call financial gravity. This is the gravity that pulls companies down into mediocrity or worse and is exacerbated by our heavily financialized economy. So to build an organization that is going to endure and is going to maintain its distinctiveness or its sovereignty over time, we have to have a force that is stronger than gravity with which we can power both the alignment that we need of people, and the structural integrity to resist outside pressure. And I call that the force of magnetic alignment. This is the mechanism by which companies gain that most valuable and underrated asset: trustworthiness. And the evidence shows that companies that have this asset, that activate this force, have numerous superpowers that conventional companies simply cannot touch.

7. Is super voting stock the silver bullet for purpose driven companies or are their other possibly better or complementary ways for purpose-driven companies to protect themselves?

It’s funny because the simple answer to your question is no. And yet I advocate for super voting shares all the time. I may be the most negative advocate of super voting shares! To understand, you have to see it this way: Imagine I went to a political science professor, an expert in political philosophy and I said, “I’m thinking of setting up a new city state, a new polis. I want your advice about what kind of governance it should have.” The professor’s going to be really excited. “Oh, great. What are you considering?” And I’ll say, “Well, I’ve only got two options. Option one is a situation in which whoever borrows the most money gets the most votes. Also, the tourists can vote, and you only have to borrow the money or be a tourist on election day, after which you can release your loans or leave the country and your vote is still binding on the whole polity.” The professor’s going to look at me and be like, “That’s pretty terrible. What else you got?” So, I’ll say, “Okay, option two is despotic emperor for life and my heirs and assigns.” The professor is going to say, “That’s all you got? Those are the only two options you can think of, really? You know, in the political science department, we’ve been working on this problem for a couple hundred years. We could maybe suggest a few other things!” That is the state of corporate governance today. It is such a paucity of thinking and originality. It is so bare of our human birthright, which is to imagine different ways that power can be shared amongst people. Human beings have been experimenting with this question since there have been human beings. So, the fact that companies are choosing despotic emperor for life to me should be read not as an endorsement of autocracy, but rather as an indictment of standard governance. Standard governance is so bad that emperor for life looks like an improvement. So yes, I do think it is an improvement. I do think there are times when that’s the best we can do, but we know from the research that it is not really the best long-term solution. We know that having too much power centralized in too few people leads to what psychologists called hubris syndrome, and many other problems besides. On top of being, ultimately not that long-term, since it’s limited by the human lifespan, this also puts a lot of founders into really an untenable and very undesirable psychological situation, where they are basically indentured servants and can never leave, for fear that their creation will be destroyed. So, maybe it’s the least bad of the current available options. But of course, we can think of far better ideas. In the book I argue for what I call “constitutional governance”, which is a set of concepts that take us beyond this false dichotomy.

8. How do you think we escape the big food doom loop? (healthy food company starts, wins customers, seeks an exit to get paid, big food makes it unhealthy and lower quality – i.e. Naked, Ben ‘n’ Jerry’s, Breyer’s, etc.)

This question is not really about food, so I’m not going to address big food. What does that even mean? Because we have a tendency to want to personalize these dramas, looking for villains. I understand that there are some villains out there. I get it. But this phenomenon that you’re describing, where someone figures out a more enlightened way to create any kind of product — doesn’t matter if it’s a food product or a tech product or a product design to bring a little beauty into people’s lives — it doesn’t matter what it is. The more successful it becomes, the more valuable it is as a target. And the more of a premium someone bigger will pay to acquire it. On this book tour, I have encountered many people who’ve told me their horror stories. They tend to want to tell food stories. That’s why I like this question. They’ll be like, look, private equity took over my favorite restaurant. Now the food is disgusting. Someone said to me a couple of weeks ago about a certain brand, “I hope they’re really successful,” and then they had to amend their statement to “Well, actually, I hope they’re somewhat successful. Successful enough to keep going, but not so successful that they get bought out by private equity.” That’s how much this idea that when things become successful, they get ruined has passed into the mainstream culture. So this is not about food. In the book, I describe this phenomenon, dating back at least two hundred years, and give the mechanics of how it happens and why. Why are we so conditioned to reenact the parable of the killing of the golden goose? And more importantly, what we can do to stop it?

9. Is it time to change the ‘corporations number one duty is to its shareholders’ narrative (aka shareholder primacy)? Is that part of what you’re trying to do with this book?

Yes. I believe that the era of shareholder primacy is actually already over, for two reasons. One is, this is an idea that has proved to be self-defeating. It was originally enacted — not in ancient times, but in the 1980s, at least in Delaware — to be beneficial to shareholders, but that is not how it has proved. We’ve actually metastasized into what I would call “extraction primacy”, in which investors themselves are now locked in a zero sum prisoner’s dilemma struggle where each has to try to squeeze as much out of everything they invest in lest someone else beat them to it. I think even investors are ready for change. The second reason I think it’s already over, and that we’re like the road runner having run off this cliff and haven’t looked down yet, is there’s a massive generational shift underway. As I mentioned before, the younger generation who has lived their whole lives under the hegemony of this idea, increasingly find it absolutely repugnant. They may not know to call it shareholder primacy, they may not realize that this is an idea that, by the way, has never been democratically enacted ever in history and therefore has no democratic legitimacy. But they are hungry for something new. And so I think our energy needs to be spent not on complaining about shareholder privacy anymore. It’s over. The question needs to be, what should the successor idea be? In the book I suggest mission primacy as one alternative.

10. You mention Novo Nordisk and its foundation in the book, which apparently is about to be passed by the OpenAI foundation for the mantle of the largest foundation (much bigger than the Bill & Melinda Gates Foundation) through their 26% ownership of OpenAI shares. Is this a model that we should encourage more startups to embrace from the outset?

I’d be very careful drawing lessons from the OpenAI experience because that company is quite singular and there’s a lot of stuff going on there quite unusual, a lot of big ego people like Elon and Sam. But interestingly, people often claim that the foundation ownership of OpenAI is unusual, and that’s not true. The idea that a for-profit company can be governed by a nonprofit foundation is an old one. The German optics company Zeiss had the structure in the 1880s. And as the question asked, Novo Nordisk has had it since the 1920s. In fact there are so many of these companies in the world that they have been studied and found to be dramatically more stable. Companies that have this structure are simply more likely to invest counter-cyclically. They are more likely to invest more in R&D. They have better financial performance and they are something like five or six times more likely to live to year fifty than conventional companies. Now the key to the structure’s stability is to have a system of checks and balances, which, as far as I understand, OpenAI struggled with for much of its existence. OpenAI had only one board, but what makes companies like Novo Nordisk, Patagonia, and Tony’s Chocolonely distinctive is that they have two entities — a for-profit board of directors who’s held accountable or in some cases even appointed by an outside board of trustees. That checks and balances, two-entity structure seems in the data to the most stable corporate form in the world.

11. As we enter the age of AI and the disruption it is beginning to cause, can the displaced really rely on enlightened capitalism to keep their families from starving?

This is a very grim question, and it presupposes one of the many, many doomsday scenarios about AI that is circulating. In order to think clearly about what it makes sense to do with AI, you have to realize two really interesting facts about this moment. The first is that almost every future scenario about this technology depends on a series of empirical facts that no one on this planet really knows the answer to. And these facts are very strange. Only a few years ago, they would have been considered post-modernist, irrelevant debates in your local philosophy department about questions like, “is there such a thing as reasoning or is it all just language?” And “what is the nature of intelligence and consciousness?” Of course, we as human beings have studied these questions for many generations. But I was on CNBC talking about this the other day — it’s rare that they are of such economic import that stock traders are wondering about them. To give one example, one of the most important questions you have to ask about AI is when or if the scaling laws will ever run out. So far, for quite a number of years,, thanks to pioneering researchers, including many far-sighted ones like my co-founder at Answer.AI Jeremy Howard, have figured out that simply by applying more computation to a very simple learning algorithm, you can create language models that seem quite intelligent, at least at first glance. So far, the more computation we use to train and run these models, the more capable they become. I think most people generally assume that this is some kind of S-curve and that eventually this curve will level off. Some even think that it already has leveled off. Others think we are years, or even decades, away from it leveling off, and of course some people believe it will never level off. This is the law of the universe. Depending on which of those things is true, the future scenarios are almost comically different from each other. A world in which the scaling laws level off next year is almost unimaginably different from one in which we have ten more years of this. And many of the doomsday scenarios, but also many of the utopia scenarios, depend critically on knowing the answer to this fundamental question about the universe that nobody knows. So, back to your question: How do we know what actions to take when the range of possible futures is so wide, so different from each other and so dependent on facts not in evidence. I think there’s only one thing that makes sense, which is to ask ourselves what are actions that would make sense, that you’ll be glad that you did, in a wide variety of potential futures? And I think that takes us out of the job of having to predict the future, which is very difficult, and rather into a more prudence-based mindset of what can be done to prepare for many possible futures. And when you go through that analysis, many of the things that you want to do to protect yourself against future AI scenarios are actually things you probably should be doing anyway. Think about having better mandatory disclosure, hardening our critical infrastructure, making sure that the gains from new technologies are widely distributed, going back to the era of widely shared prosperity. So if people are going to be displaced, should they just sit around and hope that the leaders who do the displacing will wind up being enlightened? Absolutely not. Of course not. In fact, the whole point of this book is to show how unless we make changes, the gravitational field of our financial system will warp and even destroy, turn malignant, any company. But where does the gravitational field come from? I think the most surprising part of the book for many readers is in later chapters when we reveal how the same tools that we’ve been discussing about how to create more resilient companies are also tools that can be wielded by all of us to shape the gravitational field of the future and affect what kinds of companies can and can’t form, how those companies can and cannot behave. And while some of those levers are traditional levers, like policy changes, of course., the book is primarily about the other, more surprising lovers, that I bet most readers have not thought of before.

I hope everyone has enjoyed this peek into the mind of the man behind the insightful new title Incorruptible!

Image credits: Eric Ries, Google Gemini

Subscribe to Human-Centered Change & Innovation WeeklySign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

How to Hire Like the Richest Man in the World

How to Hire Like the Richest Man in the World

GUEST POST from Shep Hyken

This article answers the question: Is it better to hire employees for attitude or for skill, especially in customer service roles?

The old saying in business, when it comes to hiring people, is this:

Hire for attitude, train for skill.

I’ve shared ideas related to this quote in several articles and videos. So, why bring it up again? First, it’s a concept worth revisiting to remind us of this important truth, especially in the world of customer service and experience. Second, I recently heard a version of this that captures the essence and further emphasizes the importance of attitude versus skill.

As I write, the richest man in the world is Elon Musk, the CEO of Tesla and founder of SpaceX, with an estimated net worth north of $500 billion. Whether you like the way he does business or not, we can’t ignore that he may have ideas worth paying attention to, and his take on this old quote is one of those ideas. The concept of hiring for attitude is driven home when he says, “Skills can be taught, but attitude changes require a brain transplant.”

Another man worth paying attention to is Jim Bush. In my book The Amazement Revolution, I interviewed Bush, who at the time was the executive VP of world service for American Express, responsible for customer support centers around the world. He shared that if he could hire someone with years of experience at a support center or working at the front desk of a hotel, he would choose the person with the hotel front desk experience.

Shep Hyken cartoon illustrating why you should hire for attitude and train for skill

Bush said, “We’re talking about human engagement, and that requires the ability to connect.” That’s why American Express began hiring people with hospitality experience. They had the attitude American Express was looking for. After being hired, they could be trained on the technical skills needed to work the computers at a contact center.

Now, before I go further, some of you might be thinking that certain jobs require specific skills, regardless of employees’ attitudes, and you are correct. A surgeon must graduate from medical school before operating. An electrician must learn the trade before wiring a home. Certain jobs require technical proficiency. However, if you hire someone with those skills who has the wrong attitude, they can harm your culture and potentially drive customers away. So, take this concept in the spirit of its meaning.

So, back to Musk’s line about attitude changes requiring a brain transplant. The comment is a bold way of saying that attitude isn’t something you can download like software. It’s hard-wired. People’s attitudes have been formed over their entire lives, from the time they were babies. Leaders who understand this focus on recruiting people who come to the job with the right mindset, with an attitude that fits the personality of the company. The takeaway is simple. Hire people who care. Then, teach them the specific skills they need to perform their job effectively. You can train for competence, but you can’t train for caring.

Image Credit: Pixabay

Subscribe to Human-Centered Change & Innovation WeeklySign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

The Fourth Inning in the Future of Work

The Future of Work Evolution: What Inning Are We In?

Editor’s Note: The State of the Game in 2026

When tech strategist Geoffrey A. Moore penned this piece in the spring of 2024, the “top of the fourth inning” was characterized by the initial, frantic rush toward generative AI adoption and a baseline shift toward customer success. Two years later, as we navigate 2026, the game has rapidly intensified.

We are no longer just talking about shifting toward “outcomes”—we are actively building the infrastructure to measure them. The baseline has evolved from simple subscription tracking to deep Experience Management Offices (XMOs) and Experience Level Measures (XLMs), proving that human-centered value is the ultimate digital metric. Furthermore, the early AI hype has matured into what we call the AI Soft Landing, where organizations are moving past experimental tools to restructure workflows around systemic, human-led collaboration. Read on to explore Geoffrey’s brilliant structural breakdown of how we arrived at this pivotal inning.

GUEST POST from Geoffrey A. Moore

It’s spring of 2024, and as Major League baseball is getting underway, everyone in tech is talking about the future of work. Let me suggest we are in the top of the fourth inning, a couple of runners on base, but still much to be decided (all with the understanding that an inning in tech lasts somewhere between one and two decades—and you thought baseball games were long!). At any rate, here’s how I see it playing out.

The first inning where tech made a definitive impact on work spanned the 1970s and 80s when the dominant paradigm was proprietary mainframe computing and the focus was on management information systems. This was an era of control cultures where the mantra was plan your work and then work your plan. IBM and Oracle were the dominant players, and workflows were organized around reports.

The second inning emerged with the rise of client-server computing in the 1990s, where the focus was on real-time business processes. This was an era of competition cultures where the mantra was give me my objectives, give me my resources, and get the hell out of my way. Microsoft and Cisco were the dominant players, and workflows were organized around documents.

The third inning emerged out of the tech bubble popping at the turn of the century, where the dominant paradigm transitioned to cloud computing combined with mobile applications, and the focus shifted from B2B complex systems to B2C volume operations. This was an era of creativity cultures where the mantra was think different. Google and Apple were the dominant players, and workflows were organized around transactions.

Now we find ourselves at the top of the fourth inning, initiated with the rise of artificial intelligence, where the focus is on as-a-service subscription business models, the economics of churn, and the importance of the customer experience. This is an era of collaboration cultures where the mantra is put customer success before everything else. The dominant players have yet to be determined, but we do know that workflows will be organized around outcomes.

And that’s the point. Information technology that began at the periphery of the business as a back office report generation utility has now migrated to the very core of the enterprise’s mission, vision, and values. That’s why digital transformation is getting so much attention. But how to transform, and how to use digital technology to ensure that customers achieve the outcomes they seek, is very much still a work in progress.

That’s what I think. What do you think?

Image Credit: Gemini

Subscribe to Human-Centered Change & Innovation WeeklySign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Top 10 Human-Centered Change & Innovation Articles of May 2026

Top 10 Human-Centered Change & Innovation Articles of May 2026Drum roll please…

At the beginning of each month, we will profile the ten articles from the previous month that generated the most traffic to Human-Centered Change & Innovation. Did your favorite make the cut?

But enough delay, here are May’s ten most popular innovation posts:

  1. Making Change Stick — by David Burkus
  2. Why You Need to Leverage Shared Values in Change Leadership — by Greg Satell
  3. Why Zero UI Will Redefine Experience Design — by Art Inteligencia
  4. Winning with Artificial Intelligence in 90 Days — Exclusive Interview with Charlene Li
  5. The Micro-Enterprise Explosion — by Braden Kelley
  6. Direction of Fit — by Geoffrey A. Moore
  7. The End of AI Data Centers — by Braden Kelley
  8. Cognitive Enhancement and the Augmented Worker — by Braden Kelley
  9. Leveraging Multi-Agent Orchestration Frameworks for Innovation — by Art Inteligencia
  10. We Must Think Less Like Engineers and More Like Gardeners — by Greg Satell

BONUS – Here are five more strong articles published in April that continue to resonate with people:

If you’re not familiar with Human-Centered Change & Innovation, we publish 4-7 new articles every week built around innovation and transformation insights from our roster of contributing authors and ad hoc submissions from community members. Get the articles right in your Facebook, Twitter or Linkedin feeds too!

Build a Common Language of Innovation on your team

Have something to contribute?

Human-Centered Change & Innovation is open to contributions from any and all innovation and transformation professionals out there (practitioners, professors, researchers, consultants, authors, etc.) who have valuable human-centered change and innovation insights to share with everyone for the greater good. If you’d like to contribute, please contact me.

P.S. Here are our Top 40 Innovation Bloggers lists from the last five years:

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Everyday Leadership

Everyday Leadership

GUEST POST from Mike Shipulski

What if your primary role every day was to put other people in a position to succeed? What would you start doing? What would you stop doing? Could you be happy if they got the credit and you didn’t? Could you feel good about their success or would you feel angry because they were acknowledged for their success? What would happen if you ran the experiment?

What if each day you had to give ten compliments? Could you notice ten things worthy of compliment? Could you pay enough attention? Would it be difficult to give the compliments? Would it be easy? Would it scare you? Would you feel silly or happy? Who would be the first person you’d compliment? Who is the last person you’d compliment? How would they feel? What could it hurt to try it for a week?

What if each day you had to ask five people if you can help them? Could you do it even for one day? Could you ask in a way the preserves their self-worth? Could you ask in a sincere way? How do you think they would feel if you asked them? How would you feel if they said yes? How about if they said no? Would the experiment be valuable? Would it be costly? What’s in the way of trying it for a day? How do you feel about what’s in the way?

What if you made a mistake and you had to apologize to five people? Could you do it? Would you do it? Could you say “I’m sorry. I won’t do it again. How can I make it up to you?” and nothing else? Could you look them in the eye and apologize sincerely? If your apology was sincere, how would they feel? And how would you feel? Next time you make a mistake, why not try to apologize like you mean it? What could it hurt? Why not try?

What if every day you had to thank five people? Could you find five things to be thankful for? Would you make the effort to deliver the thanks face-to-face? Could you do it for two days? Could you do it for a week? How would you feel if you actually did it for a week? How would the people around you feel? How do you feel about trying it?

What if every day you tried to be a leader?

Image credits: Pixabay

Subscribe to Human-Centered Change & Innovation WeeklySign up here to join 17,000+ leaders getting Human-Centered Change & Innovation Weekly delivered to their inbox every week.

Self-Acceptance Will Supercharge Your Life

Self-Acceptance Will Supercharge Your Life

GUEST POST from Tullio Siragusa

For a long time, society has demanded that we show up as good people. Do the right things and practice Godliness. The facts are that this has turned out to be an impossible expectation to fulfill. Not because we can’t be good people, and do the right things, it’s because the edict doesn’t give license to vulnerably reveal the darkness in the way of achieving the goal of being a good person.

“When we accept ourselves as a gift in the world, we begin to recognize the same in others. Whatever is external of ourselves becomes a mirror of who we are within.”

That means that if you don’t like what is external of you, simply shift what is within you.

ALL PROBLEMS SOLVED”. Simple right? Not exactly.

There is a step that most people avoid, and that is to reveal the darkness within first. At the heart of becoming the best version of ourselves, is acceptance. While historically we’ve had the pressure to always show up as if we have it all together, for fear of retribution of judgement from others, we can’ keep avoiding or masking our darkness.

It’s important to go deeper in the darkness we are in as individuals to discover the source of it, but we have to stop judging and shaming each other for being human. We are imperfect. We discover ourselves through failures., just as science discovers things through failure.

“Failure is built into the success formula of scientific discovery, it’s no different in how we discover ourselves as human beings.”

If you have darkness within you, instead of feeling shame, or guild you could shift your context and realize that our collective consciousness has chosen you to play out the darkness so you could overcome it and create the frequency for others to do the same. This is because you are the best person among all of us, to overcome it and become a beacon of Light for the rest of us.

Let me repeat that in case it hasn’t sunken in yet. YOU HAVE BEEN CHOSEN TO OVERCOME THE DARKNESS YOU ARE IN.

“The way out of hell in life… is on the other side of it. The door is just past the point of no return… only those trusting that the door is within reach, can walk through fire and gain control over everything.”

There are two ways to overcome challenges in life.

1) You work really hard to transform yourself, and to overcome the “not so good” traits; most of us end up simply suppressing who we are, but few do actually transform “some” aspects of themselves.

2) You accept yourself as you are, and you focus on becoming a being who bestows goodness in the world. When you are feeling bad about yourself, you are not good to you or anyone.

The first route will have you chasing your tail for years, and when you do fall (which happens in this imperfect reality) you’ll feel so bad, that you can’t focus on anything else. This has been the cause of depression, anger, resentment and all the chaos in the world for thousands of years. It all stems from lack of self-respect, self-love, self-dignity, self-honor, and lack of self-acceptance.

It’s impossible to accept others as they are when we still have traits, we don’t accept about ourselves. How can you accept other people’s traits, if you don’t accept yourself completely?

The second route shifts you into a parallel universe instantly, where you begin to accept others by allowing them to not be perfect, just like you.

“When you accept yourself for all of who you are, you can do the same for others, and you begin to experience life’s beauty and perfection in the imperfections.”

Acceptance shifts you into a parallel Universe where bliss is the normal mode of existence… Acceptance is being present without judgment. Having trouble with self-acceptance?

Try this simple exercise and mantra. Give yourself a hung and say:

“I am great just as I am, and I love me just as I am; I extend the same to everyone around me, and allow them to accept me as I am. I can now focus my energy on emanating the love I have for myself to the entire world and allow the world to do the same in return”.

For millenniums we’ve been going in circles feeling bad about our “character flaws”, which in some ways has kept us from achieving our greatest potential as humanity.

It’s important to get in touch with our own inner ugliness, yes… this is very important, but for no other reason than to recognize it, accept it, and find love for ourselves anyway.

“How we choose to perceive ourselves, is how we experience the entire Universe.”

Our thoughts and actions generate energy; this energy multiplies and creates a frequency for others. The more we generate the energy of compassion, love, and we shed a tear for those who suffer, the more a sense of urgency will take place worldwide to do the same.

Self-acceptance isn’t just the first step to practicing emotional intelligence, it is the way to living free of shame, and free to be our imperfect selves. My recent Rant & Grow guest, Rocky Rosen is the world’s #1 smoking cessation coach (aka the cigarette whisperer) as he turns 67 he is finally embracing self-acceptance.

Check out the coaching session with Rocky and see what commitments he makes to practice self-acceptance and supercharge his life. Maybe you’ll discover some wisdom for your own life. You can listen to the podcast right here.

Originally published at tulliosiragusa.com on September 9, 2019.

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

Managing Across Cultures

Managing Across Cultures

GUEST POST from David Burkus

If you’re leading multicultural teams, you already know that the hard part isn’t managing projects—it’s managing people. People who see work, time, leadership, and even your well-intentioned Zoom calls very differently. when your team isn’t just spread across departments or cities, but countries and cultures, those small issues can quickly snowball into trust breakdowns, missed deadlines, and a whole lot of stress.

The good news? That’s exactly where cultural intelligence comes in.

Why Most Leadership Advice Doesn’t Cut It Globally

Most leadership best practices are built on Western ideals: autonomy, authenticity, egalitarianism. And for many teams in the U.S., Canada, or Northern Europe, those principles work fine. But here’s the disconnect: over 70% of the global workforce doesn’t come from those cultures. Instead, they come from collectivist, hierarchical contexts where values like harmony, deference, and indirect communication are more important than speaking up or standing out. So, when leaders apply those Western norms across a multicultural team, problems arise. Trust breaks down. Communication stalls. Performance lags. And it’s not because the team isn’t capable—it’s because the leadership approach isn’t compatible.

That’s why cultural intelligence (CQ) is essential. According to social scientist David Livermore, cultural intelligence is a leader’s ability to recognize different cultural norms, expand their own understanding, and adapt their behavior to work effectively across those differences.

In other words, it’s not about memorizing facts like what holidays people celebrate or who bows versus shakes hands. It’s about learning to lead with flexibility, humility, and a willingness to adjust.

The Common Pitfalls of Leading Multicultural Teams

When leaders first encounter cultural differences, they often default to one of two flawed approaches: overcorrecting or oversimplifying.

Some leaders think, “Let’s celebrate every culture! Let’s learn fun facts! Let’s avoid conflict and just let people be people.” While well-intentioned, this can lead to a surface-level focus that ignores deeper dynamics.

Others take a hands-off approach: “We hired great people. Let’s let them figure it out.” But that abdication often results in misunderstandings festering until they explode—or worse, quietly eroding trust.

Then there’s the psychological safety trap. In Western teams, psychological safety often looks like open debate and direct feedback. But in many cultures, especially those where saving face is critical, this approach can feel aggressive or disrespectful.

Take Google, for example. They were early champions of psychological safety, encouraging teams to challenge ideas openly. But when they rolled out that concept globally, it backfired. Some teams became overly cautious, avoiding honesty to protect harmony. Others interpreted directness as disrespect.

The lesson? Psychological safety isn’t a universal behavior. It’s a universal need expressed in culturally different behaviors.

What Really Gets in the Way: The Hidden Barriers

To lead multicultural teams effectively, you need to recognize the specific barriers that can derail collaboration:

  1. Direct vs. Indirect Communication: In some cultures, clarity means saying exactly what you mean. In others, it means saying just enough for someone to infer your meaning. That “yes” from a team member might just mean “I hear you,” not “I agree.”
  2. Language and Fluency Gaps: When some team members aren’t fluent in the working language, it creates power imbalances. They might hold back, not because they lack ideas, but because they’re unsure how to express them. Others may interpret that silence as disengagement.
  3. Different Views of Hierarchy: In flat organizations, people are expected to challenge ideas regardless of seniority. But for team members from hierarchical cultures, speaking up—especially in front of a boss—can feel deeply uncomfortable.
  4. Conflicting Norms Around Decision-Making: Some cultures value fast, intuitive decisions. Others prefer slow, consensus-driven processes. Without clarity, this mismatch breeds frustration.

Build Cultural Intelligence with the SPLIT Framework

One of the most practical tools for building cultural intelligence comes from Harvard professor Tsedal Neeley: the SPLIT framework. It’s designed to address the core challenges of global teams—Structure, Process, Language, Identity, and Technology—and it’s especially helpful for leaders looking to lead with intention.

Structure

Structure isn’t just about org charts. It’s about perceived power. If your headquarters is in New York but your designers are in São Paulo and your engineers in Bangalore, there’s already an unspoken hierarchy. Leaders need to be intentional about flattening that perception. Reinforce that everyone’s on the same mission—different roles, same goals.

Process

Process is how you create empathy. Build in small, deliberate moments for connection. Five minutes of personal talk at the start of a Zoom call. Spontaneous Slack check-ins. And in meetings, draw out quieter voices first. Start with junior team members or those from deferential cultures. When they speak up early, it sets the tone for inclusion.

Language

Language isn’t just about translation—it’s about clarity. If some team members struggle with fluency, that’s a structural disadvantage. Set ground rules. Encourage fluent speakers to slow down and drop the idioms. Encourage non-native speakers to ask for clarification without fear. Normalize that everyone is responsible for making the conversation work.

Identity

Identity is where curiosity matters most. Don’t assume you understand what a behavior means. Ask. Learn. Invite your team to teach you about their norms—and be open about your own. The moment you switch from “leader as expert” to “leader as learner,” you earn credibility and foster mutual respect.

Technology

Technology is your connection toolkit. Use it intentionally. For trust-building, choose live video. For info-sharing, stick to well-crafted emails. And model the behavior you expect. If you ask for cameras on, turn yours on first. If you want prompt responses, respond promptly.

Cultural Intelligence Is a Leadership Discipline

Let’s be clear: cultural intelligence isn’t a checklist. You don’t become “certified” after watching one video or reading one book. It’s a leadership discipline. It’s about staying curious, adjusting your approach, and building connection—even across borders and bandwidth issues.

You’ll make mistakes. That’s inevitable. But the goal isn’t perfection—it’s progress. It’s about learning what candor means in one culture and how respect is shown in another. It’s about tweaking your leadership style not to appease, but to align.

And the result? Multicultural teams that don’t just function—but flourish. Teams where diversity isn’t a liability but a strategic advantage. Teams where trust isn’t accidental—it’s intentional.

So, if you’re leading multicultural teams and feeling a little overwhelmed, take a breath. Start small. Ask better questions. Listen a little longer. And lead a little differently.

Because cultural intelligence isn’t just the key to global collaboration. It’s the new core competency for leadership.

Image credit: Pexels

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.

The Anatomy of Agentic Trust

A Mechanistic Interpretability Framework for Change Leaders

LAST UPDATED: June 5, 2026 at 3:13 PM

The Anatomy of Agentic Trust - A Mechanistic Interpretability Framework for Change Leaders

GUEST POST from Art Inteligencia


The Impasse of the Black Box: Why Agentic AI Demands a New Trust Paradigm

Digital transformation has reached an inflection point. Organizations are moving away from traditional, deterministic software and basic copilots toward Agentic AI—autonomous systems capable of executing complex, multi-step operational workflows with minimal human oversight. While this shift promises unprecedented efficiency, it introduces a severe psychological and operational barrier: The Wall of Trust.

The Shift to Autonomy

Unlike previous iterations of artificial intelligence that relied on simple pattern-matching or isolated text generation, agentic systems possess agency. They can formulate plans, interact with external software ecosystems, and make consequential business decisions independently. However, because these systems are built on top of massive deep learning architectures, their reasoning remains entirely opaque.

The Psychological Friction of Current AI Explanations

Traditional approaches to Explainable AI (XAI)—such as post-hoc approximations, saliency maps, or text-based self-justifications—are no longer sufficient for enterprise governance. These methods merely show what data correlated with an output; they do not reveal the actual underlying computational logic. When an autonomous agent makes a flawed decision, a post-hoc explanation acts as a guess rather than an audit trail. For a workforce tasked with collaborating alongside these machines, this lack of transparency breeds deep-seated skepticism.

The Change Management Mandate

Successful innovation and experience design depend entirely on psychological safety. Change leaders cannot integrate autonomous agents into hybrid human-machine teams if the machine’s logic remains inscrutable. To transition employees from defensive resistance to confident collaboration, organizations must establish absolute legibility. Mechanistic interpretability provides the exact verifiable transparency required to align AI agents with human ethics, compliance mandates, and organizational values.

Demystifying Mechanistic Interpretability: From “Black Box” to Open Circuit

To dismantle the black box, innovation and change leaders must embrace a paradigm shift in how we audit artificial intelligence. Mechanistic Interpretability (MI) moves away from treating neural networks as abstract, unknowable minds. Instead, it approaches them like complex, physical objects—akin to an intricate mechanical watch or an integrated circuit board—that can be systematically disassembled and reverse-engineered.

The “Neuro-Industrial” Approach

Rather than merely observing what goes into a model and what comes out, MI focuses on internal computational mechanics. By treating deep learning structures as physical systems waiting to be mapped, researchers and engineers can trace the exact pathway information takes as it moves through the network. This shifts the conversation from passive observation to rigorous, empirical auditing.

Deconstructing the Neural Architecture

Understanding this open-circuit paradigm requires looking at three core components of modern model architecture:

  • The Communication Channel (The Residual Stream): Think of the residual stream as the primary information highway of a Large Language Model. As data passes from layer to layer, each computational mechanism reads from and writes to this central highway, iteratively refining the concepts the model is processing.
  • The Challenge of Superposition: Deep learning models are incredibly efficient compactors. Through a phenomenon known as superposition, a network can compress thousands of overlapping concepts into a relatively small number of neurons. This results in “polysemanticity”—where a single neuron might fire for a medical diagnosis, an ancient historical event, and a specific lines of code, making raw network readouts look like total gibberish to humans.
  • The Solution (Sparse Autoencoders): To untangle this mess, researchers use an auxiliary tool called a Sparse Autoencoder (SAE). The SAE acts as an analytical lens, expanding the compressed neural activity back out into an uncompressed, highly specific map of distinct business concepts and features. Polysemantic neurons are separated into clean, human-readable concepts.

Mapping the Circuits

Once the concepts are isolated by Sparse Autoencoders, change and safety leaders can trace how individual components connect to form causal, end-to-end pathways—or circuits. These circuits execute specific pieces of logic, such as a circuit that detects tax compliance rules or a circuit that handles data privacy boundaries. Mapping these circuits turns an opaque mathematical matrix into a transparent, visual map of organizational logic.

The Commercial Frontier: Leading Organizations and Startups Shifting MI from Theory to Tooling

What began as an academic and safety-centric pursuit has quickly evolved into a critical layer of the enterprise AI value chain. As organizations demand verifiable trust before deploying agentic workflows, a robust commercial ecosystem has emerged. Today, the development of Mechanistic Interpretability tools is divided among frontier research labs, open-source consortia, and specialized AI safety startups.

Frontier Research Labs: Setting the Scale

The foundational model developers themselves are treating internal architectural translucency as both a primary safety barrier and a competitive advantage.

  • Anthropic: Widely recognized as a pioneer in dictionary learning, Anthropic demonstrated commercial-scale concept mapping by isolating millions of abstract, safety-critical, and real-world features inside its Claude models. Their pioneering work in circuit tracing maps not just which features are active, but how they causally influence each other in sequential processing chains.
  • OpenAI: Operating at massive computational scale, OpenAI has focused on automating the interpretability pipeline itself. By utilizing advanced Large Language Models as automated “feature explainers,” they systematically analyze, score, and catalog millions of dense neuron activations simultaneously across models like GPT-4, laying the groundwork for algorithmic “lie detectors” built directly into model internals.
  • Google DeepMind: DeepMind significantly accelerated industry-wide adoption with the release of Gemma Scope, a massive, comprehensive open-source interpretability toolkit mapping across the entirety of its Gemma model families. This initiative effectively democratizes MI, giving enterprise change and innovation leaders the open tools needed to audit fine-tuned models independently.

Open-Source Consortia

Bridging the gap between frontier research and accessible development is EleutherAI. Through specialized open-source libraries like sparsify, EleutherAI provides researchers and enterprise engineers with the standard blueprints required to train Sparse Autoencoders (SAEs) and transcoders directly on HuggingFace transformers, allowing organizations to extract custom, localized operational feature dictionaries without relying on proprietary third-party APIs.

The Emerging AI Governance & Steering Startup Ecosystem

As the market shifts from post-hoc model analysis to real-time behavioral intervention, a specialized group of AI safety, security, and compliance startups has emerged. These early-stage innovators are building platforms that operationalize MI principles for the enterprise:

  • Algorithmic Auditing & Protection Platforms: Emerging vendors—including teams like Protect AI, Turing, Holistic AI, and Enkrypt AI—are actively developing continuous monitoring guardrails, neural audit logs, and PII containment shields.
  • From Observation to Intervention: Rather than just notifying a business that an autonomous agent has hallucinated, the vanguard of this ecosystem is building enterprise toolsets focused on feature steering. By giving compliance officers and change managers the ability to programmatically clamp down or amplify specific feature vectors, these platforms provide an exact knob to safely steer agent behavior in production environments without requiring costly model retraining cycles.

The Collaborative Interface: Designing the Human-Machine Audit Trail

For change and innovation leaders, a technical map of a neural network is only useful if it can be translated into operational reality. To turn Mechanistic Interpretability from an engineering luxury into a practical governance mechanism, organizations must implement a standard action loop. This practical paradigm is defined by three continuous operational steps: Locate, Steer, and Improve.

1. Locate (The Diagnostic Phase)

When an autonomous AI agent produces an unexpected anomaly, drifts from compliance, or triggers a customer experience failure, traditional troubleshooting is useless. Under the MI framework, operations teams initiate the Locate phase. By utilizing Sparse Autoencoders, corporate compliance teams can systematically look under the hood to isolate the exact subgraphs and internal feature nodes that dictated the agent’s flawed decision path. Instead of guessing why an error occurred, leaders can pinpoint the specific computational circuit responsible for the behavior.

2. Steer (The Real-Time Intervention Phase)

Once a problematic circuit or feature node is located, the organization does not need to undergo a weeks-long, financially draining model-retraining process. Instead, leaders use feature steering to intervene directly. By programmatically adjusting, clamping, or dampening specific feature activations within the live system, operations teams can instantly align the agent’s behavior. For example, if an insurance agent begins using unapproved geographic criteria to assess risk, a compliance manager can safely dial down that specific feature vector without degrading the agent’s overall processing capabilities.

3. Improve (The Continuous Alignment Phase)

The final phase transitions the organization from reactive intervention to proactive refinement. Over time, data engineers, risk managers, and business unit leaders iteratively review the agent’s global modular vocabulary. By continuously updating and refining these feature dictionaries, the enterprise can permanently align autonomous workflows with changing regulatory landscapes, ethical guidelines, and internal corporate values. This creates a living, transparent human-machine audit trail that ensures autonomous systems remain accountable to human intent.

The Human-Centered Angle: Using Circuit Translucency to Drive Adoption

The ultimate success of any digital transformation initiative hinges on the psychology of the people expected to drive it. Technology alone does not yield ROI; adoption does. By turning the “black box” into a translucent, auditable map of circuits, Mechanistic Interpretability addresses the deepest root cause of workforce resistance: the fear of the invisible, unaccountable driver.

Abolishing the “Us vs. Them” Dynamic

When autonomous agents are introduced as inscrutable forces that magically output decisions, an adversarial dynamic inevitably forms between employees and technology. Teams view the AI as an opaque competitor designed to replace or undermine their judgment. Providing an interactive, auditable look “under the hood” radically reframes this relationship. When employees can visually trace the model’s logic pathways, the AI shifts from a mysterious threat to a legible, controllable tool. Demystification actively dissolves defensive skepticism and replaces it with shared ownership.

Designing the Experience of AI Auditing

Innovation and experience design leaders must proactively design the workflows that connect humans to these neural circuits. This requires upskilling traditional Subject Matter Experts (SMEs)—such as underwriters, clinicians, or compliance officers—from passive users into active “circuit overseers.” Instead of forcing SMEs to learn complex linear algebra, organizations must build intuitive, human-centered dashboard experiences. These interfaces translate complex Sparse Autoencoder feature dictionaries into plain language, empowering business leaders to confidently monitor, validate, and sign off on automated reasoning.

The Safety-Trust Horizon

Psychological safety cannot coexist with unpredictability. True confidence is built on empirical predictability—knowing exactly where the guardrails are and how to enforce them. By establishing a verifiable baseline for risk mitigation, circuit translucency gives operations teams the concrete evidence they need to trust autonomous systems. When a team knows they can structurally audit a workflow, catch compliance drift before it impacts a customer, and pinpoint exactly why an anomaly occurred, they can deploy agentic workforces at scale with absolute confidence.

Operationalizing the Framework: A Roadmap for Innovation Leaders

Transitioning an organization from opaque, unverified AI deployments to a translucent, mechanistically interpretable architecture requires an intentional, staged approach. Innovation and change leaders cannot implement this infrastructure overnight. Instead, they must systematically align technical capabilities with human experience design. This roadmap provides a practical three-phase deployment strategy to operationalize agentic trust across the enterprise.

Phase 1: Diagnostic Readiness and Risk Mapping

The first step is identifying high-stakes operational workflows where opaque agent logic presents an unacceptable risk to compliance, organizational stability, or brand trust. Leaders must audit their current AI roadmap and pinpoint “red zone” processes—such as autonomous financial underwriting, automated contract enforcement, or clinical triage routing. By scoring these workflows based on regulatory exposure and the psychological impact on the employees overseeing them, organizations can prioritize exactly where mechanistic transparency is required to maintain operational stability.

Phase 2: Architectural Translucency and Feature Extraction

Once high-risk workflows are mapped, innovation leaders must partner directly with AI engineering and data science teams to build out the technological transparency layer. This phase involves integrating open-source frameworks or commercial governance platforms directly into fine-tuned enterprise models. Engineers deploy Sparse Autoencoders (SAEs) and transcoders across the model’s layers to untangle polysemantic neurons, systematically extracting a structured, human-readable dictionary of the specific business concepts, compliance rules, and operational parameters the agent uses during execution.

Phase 3: Cultural Integration and Co-Creation Loops

The final phase embeds this structural transparency directly into the company’s operating model and culture. Change leaders must design and establish cross-functional governance loops where compliance officers, risk managers, change management practitioners, and front-line business leaders systematically review and steer agent behavior. By designing intuitive dashboards that translate extracted features into plain language, organizations empower non-technical personnel to participate in feature-steering exercises, transforming AI alignment from a back-office engineering chore into a collaborative corporate discipline.

Conclusion: The Future of Co-Elevation

As organizations stand on the precipice of widespread Agentic AI deployment, a critical truth becomes apparent: the ultimate bottleneck to scaling artificial intelligence is not computational power, data density, or algorithmic sophistication—it is human trust. Businesses cannot capture the exponential ROI of autonomous workflows if their own teams pull back in skepticism, or if compliance frameworks reject the inscrutable nature of the systems driving them.

The Core Philosophy

Mechanistic Interpretability represents far more than a technical patch for AI safety. It is a fundamental philosophical shift that treats neural networks with the same empirical rigor we apply to physical engineering. By transforming the “black box” into a legible blueprint of interconnected circuits, we strip away the unhelpful mystique surrounding deep learning. This structured transparency provides the absolute bedrock for psychological safety, transforming autonomous agents from opaque wildcards into predictable, reliable partners.

The Innovation Call to Action

Forward-thinking innovation and change leaders must stop viewing AI safety and interpretability as a narrow, back-office technical function left solely to data scientists. True, sustainable digital transformation requires a holistic approach. It is the responsibility of culture builders, experience designers, and corporate strategists to champion architectural translucency. By operationalizing Mechanistic Interpretability, enterprises can successfully bridge the cognitive divide, mitigate systemic operational risk, and unlock the true potential of a highly confident, collaborative, and co-elevated human-machine workforce.

Frequently Asked Questions

To help both your human teams and automated search crawlers understand the intersection of AI safety and organizational change, this section includes a standard human-readable FAQ alongside a structured JSON-LD Schema block optimized for modern answer engines.

1. How does Mechanistic Interpretability differ from standard Explainable AI (XAI)?

Traditional Explainable AI (XAI) usually generates post-hoc guesses or approximations—like text descriptions or heat maps—of why a model arrived at an output. It tells you what inputs correlated with the result, but not the actual path taken. Mechanistic Interpretability (MI) reverse-engineers the network itself, unpacking compressed neural activity to reveal the literal computational “circuits” and logical workflows inside the model. It moves from correlation to true mechanical causation.

2. Why is structural transparency critical for human-centered change management?

Successful digital transformation requires psychological safety. When organizations deploy fully autonomous “Agentic AI” workflows without visibility, employees experience defensive skepticism because they cannot audit, predict, or trust the system’s logic. By making the model’s internal reasoning translucent, change leaders can transition human teams from resistant onlookers to confident collaborators who can proactively steer and manage their AI partners.

3. What is “feature steering” and how does it protect an organization?

Feature steering is the ability to programmatically amplify, clamp, or dampen specific concept vectors isolated inside a model using Sparse Autoencoders (SAEs). Instead of undergoing a long, expensive retraining or fine-tuning process when an AI agent drifts out of compliance or experiences a workflow anomaly, compliance and innovation managers can adjust the model’s specific internal logic dials in real time to ensure safe, ethical execution.


Disclaimer: This article speculates on the potential future applications of cutting-edge scientific research. While based on current scientific understanding, the practical realization of these concepts may vary in timeline and feasibility and are subject to ongoing research and development.

Image credits: Gemini

Subscribe to Human-Centered Change & Innovation WeeklySign up here to get Human-Centered Change & Innovation Weekly delivered to your inbox every week.