Re:Wire Ep. 04 – Deliberation Debt, Leadership Fluency and the Human Layer of AI with Niketa Jhaveri

Re:Wire, the Experience Haus podcast Episode 04

Host: Amit Patel, Founder, Experience Haus
Guest: Niketa Jhaveri, UX Leader at Amazon; Founder of Nivan

Watch: https://www.youtube.com/watch?v=viz6LyKHlJY
Listen: https://open.spotify.com/episode/0oVNzyCgVRW7Gc8CYrTFDR?si=2WaGjkXcQjaiWACZ7KZCcA

In this episode of Re:Wire, Amit sits down with Niketa Jhaveri, a UX leader at Amazon and founder of Nivan, an innovation conference exploring how fast-moving AI is reshaping people, culture and leadership.

The conversation covers why leadership in the AI era is a people problem before it’s a technology problem, the three things Niketa believes every leader needs to build, fluency, judgment and context, her concept of “deliberation debt” and why moving fast without reflection creates a cost you can’t undo, what actually needs to change about organisational structure as agentic workflows take hold, and why cultural intelligence matters more, not less, as AI starts having the conversations with customers that used to be human. It closes with her view on how a leader’s job itself is changing, now that answers are cheap and trust is the scarce thing.

Please note, this transcript has been lightly edited for clarity and length.

Introducing Niketa and Nivan

Amit: Welcome back, everyone, to this latest episode of Re:Wire by Experience Haus, where we explore everything happening in the world of AI, leadership and transformation, and have great conversations with people in the industry. I’m going to skip my usual introduction this time and let Niketa talk about herself. Niketa, we met in New York a couple of months ago, hit it off straight away, and I’ve been looking forward to this ever since. Give us a hello and an introduction.

Niketa: Thanks for having me, Amit, I really appreciate it, and I love what you’re doing with Re:Wire, I’ve listened to all the episodes. I’m Niketa Jhaveri, I’m a UX leader at Amazon, and I also host an innovation conference called Nivan. The idea behind it is that AI is moving fast, and I wanted a way of bringing people together to talk about how that speed actually affects people, culture and leadership.

Amit: Tell us a bit more about Nivan. You had your first one earlier this year, and there’s more coming up, right?

Niketa: Yes, a few things happening there. We ran our first ever Nivan in May, in Chicago, nineteen speakers, fifteen partners, one unforgettable day, with people coming in from Netflix, Microsoft, Amazon and other great companies. The format’s a little unusual, I call it time travel: the morning is the past, midday is the present, and the afternoon is the future. I fundamentally believe that even as new technology and new horizons show up, the fundamentals from past innovation shouldn’t be forgotten, they should be carried forward. We also run a format called Creative Duel, two people on stage professionally disagreeing with each other, debating things like human versus AI, scale versus craft, personalisation versus privacy. That format got popular enough that we’re doing a standalone evening version of it in Chicago on September 24th, dinner by the lake, wine and beer, and we’re taking the same format to Seattle in November. The flagship event comes back next May, in 2027.

What Leadership Looks Like in the AI Era

Amit: What you just described, past and present sitting alongside each other, is a nice way into where I want to take this. At Experience Haus we’re working with organisations trying to plot out their North Star workflow design, and that always requires real leadership and a human reorganisation alongside it. What does leadership actually look like to you right now? Any core principles standing out?

Niketa: You’ve basically answered your own question by saying “AI era,” that’s exactly where I was going to go. With so many AI tools available, accelerating so much, that’s a good thing. But I think leadership right now is a people problem before it’s a tech problem. We’re investing heavily in AI, in tooling, and rightly so, but leadership isn’t just about the technical side, it’s about people and culture, and that’s what shows up first, before anything else does. I read that ninety-three percent of Fortune 100 data leaders name culture and change management as their top barrier, only seven percent name technology. So yes, adopt the technology, but make sure your team, your people and your culture stay intact through it.

Amit: We’ve seen that play out directly. We’ll do great work helping a team plot their North Star, the workflow, the agentic approach, and then the moment it comes to actually implementing it, the people and the culture become the blocker. What advice would you give someone navigating that right now?

Fluency, Judgment, Context

Niketa: I’d give three pieces of advice. Number one is fluency. Understand the skills, understand where things are going, what tools and data people are actually using. You don’t have a choice here, you have to learn it, and there are plenty of resources out there, Experience Haus being one of them. That’s the basic level, fluency.

Number two is judgment. I know it sounds strange to say “learn judgment,” but can you tell when your data is showing you something wrong, when the AI isn’t actually giving you the right result? At Amazon we talk about needing to “be right, a lot.” You need the judgment to know whether the AI you’re executing, the code it’s generating, or a decision it’s informing, is actually correct.

Number three is context. Can you tell when something’s right, but still wrong for your situation? That’s where context comes in. Most companies today only really invest in the first layer, fluency, learn, learn, learn. But how do we make sure people are also developing judgment and context? You might be measuring usage because a customer interacted with your AI tool, but what if the decision they actually made was not to use it, because they judged it wasn’t the right product for them? You need all three layers working together.

Amit: That makes me think leaders have to be far more hands-on than they’ve historically been. Is that the direction you’re pointing at?

Niketa: Yes. Leaders need to be hands-on because they themselves have to evaluate whether their team’s product is actually right for the customer experience they’re trying to build. But I also want to introduce a model I’ve been circulating, deliberation debt. It’s the idea that the faster you go, the slower you actually end up being, it’s the cost of deciding faster than the question deserves, and that interest comes due later. I’ve written about this on my website, I’ll send the link round. Yes, leaders have to learn the tools, but they also need to understand what deliberation debt they’re taking on. Everyone wants to move faster, and that’s fine, but can you pause for a second and ask what you might be missing? Because once a product is released, you can’t go back and fix that debt, you won’t even see it until much later.

Ownership from the Bottom Up

Amit: That makes sense, and it connects to something that came up a lot when we met at a conference in New York, the question of who owns it when something goes wrong. Leaders having to step up and own whether they were actually hands-on or not.

Niketa: I think the responsibility sits with everyone using the AI tool, whether they’re building something or making a decision with it. Even at the software engineering level, the people coding, the responsibility for the customer outcome doesn’t disappear just because it sits with senior management too. Responsibility runs from the bottom of the organisation to the top. That’s how leadership is changing, and so is the culture around it.

Reimagining the Organisation for Agentic Workflows

Amit: I want to pivot slightly to the human, organisational lever. As agentic workflows kick in and we see more autonomous work happening, what happens to the organisation underneath it? I’m actually writing about this in tomorrow’s Rewired issue, because it’s one thing to define your North Star and figure out which agents you need, but what happens to the structure around it? What does leadership need to do to reimagine what their organisation looks like in a world of agentic workflows?

Niketa: This is an interesting one. First, leadership needs to invest in their existing people, make sure everyone’s trained properly and getting the right skills. I work inside a design and research organisation, and honestly, I think it’s less about changing the org structure itself and more about investing in people and skill sets to make things efficient. That said, you are starting to see new roles open up, hiring for AI-specific scientists, for instance, because the old skill set in a given department no longer matches the product being built. Data literacy is one example. For those of us in UX and design, who think of ourselves as creative people focused purely on design, data literacy might sound irrelevant. It isn’t. You need to be able to read whether what the AI is producing, and how customers are actually interacting with it, makes sense.

The second skill set is workflow automation. Automation is everywhere now, you can ask an AI tool to go and create a viral video for you, which might mean you don’t need as large a marketing team to execute it. But you still need people who can judge whether that automation is actually having the intended effect.

The third, and the most important, is governance. Who is governing whether the AI is making the right decisions for you? Is customer data protected, is there real privacy around it, what happens if the AI pushes data somewhere it shouldn’t? Organisations need all three of these skill sets built in, whether that’s upskilling existing people or creating new roles for it.

Unconscious Incompetence and the Cost of Speed

Amit: What I love about that is it echoes something we worked through in a workshop last week at a local university, three layers: intent, intelligence, and something like an ethics or integrity layer. When we evaluated existing startups already in market, they weren’t failing at just one of those layers, often several at once. It made me realise that the big thinking, the right questioning, is getting skipped simply because execution has become so fast. We can ship things now, but nobody’s gone back to properly evaluate them.

Niketa: Exactly, and that’s exactly where deliberation debt comes in, you’re moving at a speed you’re not actually ready for, and that’s where the debt accumulates. It ends up more costly than the interest on a mortgage, because you can’t go back and undo it.

Amit: That’s making me think the faster you move, the bigger the risk, it’s basically walking away from fundamental user research, which exists to reduce the chance of getting it wrong. But because execution is so fast now, we’re just running ahead at speed and skipping it. Provocative question: do you think people are actively choosing to ignore those questions, or are they just not equipped to ask them?

Niketa: I’d frame it as unconscious incompetence, not conscious avoidance. People don’t know what they don’t know, and they’re moving at the speed everyone around them expects, their customers, their leadership. It’s not really anyone’s individual fault. But if you imagine it’s 2030, and I do love this time travel framing, we’ll be sitting on top of millions of lines of code that need revisiting because something broke and nobody can quite figure out how to fix it. That’s deliberation debt, the direct cost of trying to go faster than the thinking could actually support.

Closing the Skills Gap

Amit: I know we’re tight on time, but I want to go back to the skills gap you mentioned, because it ties directly into organisational structure. What does a team need to look like once this work has been done, and what advice would you give leaders who are recognising they have a skills gap right now?

Niketa: First, invest twenty percent of your people’s time into getting genuinely fluent, learning, continuously. Every single day a new AI tool shows up doing something slightly different, and you need to actually know it to build the right product. Second, adopt tools and judge for yourself whether they’re right for you, rather than assuming something’s right just because a competitor is using it. That’s not necessarily true for your environment or your customers, and that’s exactly where judgment comes in: is what it produces actually the experience you want, and will it hold up, will it scale and protect you, ten years from now, not just in this moment?

The third thing I’d highlight is culture, and I don’t mean internal company culture, I mean the culture of the markets you actually serve. Companies building for global markets often build something for the US or UK and assume it’s right for every customer in the world. Cultural intelligence matters here, and it isn’t a new idea, we’ve talked about this for years. What’s new is that AI is now the one having the conversation with your customers, in Brazil, Mexico, India, wherever, so the question becomes whether AI is actually adapting to the norms of those markets, not just translating language. So overall, it comes back to building these underlying skill sets and knowledge yourself, rather than just leaning entirely on AI. AI will make you faster, one hundred percent true, just make sure it doesn’t make you dumber in the process.

Amit: That resonates completely. I made a similar point in a talk a couple of years ago, that cultural intelligence needs to sit alongside emotional intelligence and empathy in any team. None of this is new, these are products and ideas that have existed for decades. What’s changed is that AI has sped up how quickly a product can go global, but the fundamental thinking around adapting to context on a global scale hasn’t changed at all, we’re just now feeling the pace of it.

Answers Are Cheap, Trust Is the Job Now

Niketa: Exactly. I think our job, leaders like you and me, used to be having the answer ready. Now, answers are cheap, you can get one from a single line of a prompt. What the leader’s job actually is now is asking whether that answer is trustworthy, whether you can rely on it to drive the right decisions for your team or organisation. The leader’s role has shifted, it’s not about needing more answers, it’s about bringing the right judgment to the answers you already have.

Amit: That’s such a perfect way to close this out. Before you go, tell people where they can find out more about the conference, and about you.

Niketa: For the conference, go to nivan.live. And if you’d like to read more on deliberation debt, cultural intelligence and the other things I write about, always through that time travel lens, go to niketajhaveri.com, you’ll find it all there.

Amit: Amazing, and when you’re in London in November, let’s do an in-person part two. This has been absolutely brilliant, thank you so much for joining us.

Niketa: Absolutely, I’d love that. Thanks for having me, it’s been an honour.

Niketa Jhaveri is an UX leader at Amazon and founder of Nivan. Follow Re:Wire for more conversations on leadership, unlearning and the future of work.