Why the AI era demands more from product managers, not less | Perk President of Product

Nikita Miller spent 15 years building products before becoming President of Product and Technology at Perk, an AI platform for business travel and spend management. She made the move from the U.S. to Barcelona about a year ago, arriving right in the middle of Perk’s transition from a single travel product to a multi-product platform. The timing meant navigating that transformation while also rethinking how product teams work in an era where AI has changed almost everything about how work gets done. In this episode, host Cassio Sampaio sits down with Nikita to talk about why the AI era demands more from product managers, not less, the UX renaissance she sees happening in B2B, the concept of shadow work and why removing it is one of the biggest product opportunities right now, and what she tells early-career product managers who are trying to figure out where to invest in themselves when the landscape keeps shifting.

Subscribe to the Product Talk podcast on Spotify and Apple Podcasts and catch every conversation with leading product executives.

---

---

LISTEN ON ITUNES

---

LISTEN ON SPOTIFY

---

---

Show Notes

1. Nikita joined Perk about a year ago, arriving right in the middle of the company’s transition from a single business travel product to a multi-product platform that now spans travel, spend management, events, and several products in incubation. She arrived expecting the biggest adjustment to be cultural, moving from U.S. product culture to a company headquartered in Barcelona with over 60 countries represented on the team. What actually surprised her most was the ambition of the Perk team, which she describes as just massive.
2. The biggest single change in how she runs product teams is AI, not geography or culture. Fifteen years of building taught her how to structure cross-functional teams around product, design, and engineering. What AI has done is shrink those teams while expanding what each person can do. She has designers doing more front-end coding, engineers getting closer to UX, and product managers moving faster across more parts of the build cycle. The crafts are still distinct and still matter; the people within them are just flexing into new areas.
3. She frames this moment as a once-in-a-career opportunity, and it is one of the few things she sees as genuinely independent of where you are in the world. Whether you are in the U.S. or Europe or anywhere else, every team building software products is navigating the same disruption at the same time with the same incomplete playbook. There is something equalizing about that, and she finds it energizing rather than unsettling.
4. Perk’s original wedge was business travel management for small and medium-sized businesses, and the word massive is intentional when she describes it. Business travel in a B2B context is not just booking flights and hotels. It includes expense policies, duty of care, CFO-level spend controls, approval workflows, per-country regulatory complexity, and real-time reconciliation. Winning in that space required building genuine local expertise across every European market before expanding into the U.S., where the context is different again.
5. Travel was the beginning because travel is just one way employees spend money on behalf of a company. Once Perk owned that relationship, the adjacency to expense management was natural: every flight needs approval, every trip generates receipts, every receipt needs to be reconciled. Launching spend, and then cards, and then events is not a pivot; it is the same user base and the same financial workflow extending further into the company’s operations. One in four U.S. companies still manages all of this manually without a spend management tool. That is the opportunity.
6. The UX renaissance she sees in B2B is real, and it has been building for a while. Business travel users book personal travel all the time on consumer apps, and they carry those expectations into the tools their employer gives them. The speed, the personalization, the quality of recommendations: those are not enterprise-versus-consumer distinctions anymore. They are just baseline expectations, and B2B products that do not meet them are increasingly vulnerable.
7. The administrator and manager experience is just as important as the end user experience at Perk, and it follows the same principle. Approving a travel request or keeping employees within policy should not require navigating a complex reporting interface. The expectation is that you can ask a question in natural language and get a clear answer. That is the bar now, and it has raised the importance of design and front-end craft at a moment when some in the industry had been questioning whether those roles still mattered.
8. On whether AI helps disruptors or incumbent giants, she gives a nuanced answer. The instinct is to say that AI levels the playing field, but she thinks the opposite might also be true: incumbents have the capital to move aggressively on AI, and AI might actually enable them to simplify and segment their products in ways that were not previously possible. The tool is available to everyone. The question is who gets there faster, and that is not always the smaller, newer player.
9. Perk is actively working to disrupt itself. Rather than waiting for a competitor to use AI to undercut what they have built, they are running a parallel stream of work asking what the product could look like if they were building it from scratch today. She describes this as something the American in her is wired to do: assume someone is always behind you trying to get there, and use that as a reason to move faster.
10. Her preferred term for what many people are calling taste in the AI era is discernment. The distinction matters to her. Taste suggests aesthetic judgment. Discernment is the ability to judge which problems to solve, in what order, with what urgency, given all the complexity and competing signals that a product leader has to navigate. You can build almost anything now, which means the value is not in the ability to build. It is in knowing what is worth building.
11. Discernment is not innate and it is not something that AI has replaced. It is a craft that is trained over time through experience, data, customer depth, and strategic clarity. The articles predicting that product managers and designers would be obsolete by 2027 frustrated her not just because they were demotivating but because they were missing this entirely. The judgment about which problem to solve and why now is still human, still hard, and still the most important thing a product leader contributes.
12. The concept she and her team use to frame product opportunities at Perk is shadow work: all the rote, manual, cumbersome tasks that customers are doing right now that could be automated or handled by AI. The goal is to identify that shadow work in customer workflows and remove it, freeing customers to apply their own discernment to the things that actually require human judgment. It is a useful frame because it separates what should be automated from what should not, and it makes the product opportunity concrete.
13. On the question of whether AI will eventually become the de facto interface for everything, she is honest that she does not know. What she does know is that adoption is varying widely right now. Perk has an MCP so customers can book travel and manage expenses through AI tools directly, and some customers are enthusiastically using it. Others, particularly those outside the tech industry, are not interested yet and may not be for a while. Meeting customers where they are is the operating principle, not betting on a single interface future.
14. She also pushes back on the idea that all interfaces will converge on one modality, whether voice or chat or anything else. People think differently: some in language, some in images, some through spatial or visual interfaces. A one-click travel booking that works because Perk already knows everything it needs to know about you is a very different interaction than a voice conversation with an AI. Both might have a place, and she is skeptical of anyone who confidently predicts only one survives.
15. The advice she gives early-career product managers has not changed much despite everything else that has: read more. The core skill of product management, now more than ever, is the ability to understand a problem clearly and communicate it clearly, whether to a human audience or through the prompts and systems you build for AI. That clarity of thinking and expression is developed through reading broadly, absorbing ideas from outside your immediate domain, and building a richer mental model of how things work.
16. She connects this to what she thinks of as classical education: there are foundational things about learning, acquiring knowledge, and developing understanding that do not become obsolete with each new wave of technology. The ability to reason clearly, to articulate a problem with precision, to distinguish signal from noise: these are persistent skills. The tools change. The underlying capability compounds across every tool change.
17. The international team at Perk, with over 60 countries represented, has taught her that cultural navigation is its own craft. European and American teams often have different approaches to directness, risk tolerance, and pace. Getting that to work together without losing the best of either requires intentional investment, and she finds it one of the genuinely joyful challenges of her current role even on the days it is difficult.
18. On the deeper question of what remains irreducibly human as AI becomes more capable, she is clear about what she does not want to give up. She wants to hug people. She wants long meals at a dinner table with other humans. Perk is in the business of getting people to places where they can be together in person, and that mission feels more important to her, not less, as AI handles more of the logistics around it.
19. She is fundamentally an optimist about the AI era, and she thinks you have to be if you are going to stay in product. There is always more to learn and more to figure out, and she finds that far more interesting than the doomsday framing that periodically dominates the conversation. Investment keeps flowing, productivity is increasing, and the people who are learning to create and direct agents are building new skills rather than losing old ones.
20. Her parting thought on what early-career PMs should hold onto: optimism is not naive in this field, it is professional. Product management exists to identify real problems, make sound judgments about which ones are worth solving, and build things that create genuine value for the people who use them. That has not changed. What has changed is the surface area of what is buildable, which makes the judgment more important, not less.