CPO Rising Series: Rakuten Rewards CPO on Human Judgment in the AI Era

What does it mean to govern a decision model instead of a roadmap? In this episode of the CPO Rising Series hosted by Products That Count Resident CPO Jay Patel, Rakuten Rewards CPO Nilesh Khandelwal will be speaking on human judgment in the AI era and how AI is shifting the product leader’s job from execution management to decision architecture. He also shares his three-bucket prioritization framework, the concept of “glues” as a retention lever, and why saying no to guaranteed revenue is sometimes the most strategic move a CPO can make.

Join us for new conversations with leading product executives every week. Roll through the highlights of this week’s event below, then head on over to our Events page to see which product leaders will be joining us next week.

Show Notes:

1. As execution becomes cheaper and faster, human judgment becomes more valuable, not less. The linear relationship between team size and impact is broken. A smaller team with the right decision framework now outperforms a larger team without one. The CPO’s job is no longer to drive output — it is to govern the quality of decisions that determine whether the company is on a positive trajectory.
2. Bad decisions get accelerated by AI just as fast as good ones. The same tools that compress timelines and multiply output will amplify poor framing at scale. Getting the problem definition right before a single line of code is written is now more consequential than it has ever been, because the cost of a wrong direction compounds faster.
3. Product management is transitioning from writing requirements to governing a decision model. The future role of product is not to produce specs but to maintain the right framework, the right guardrails, the right data, and the right checkpoints where human judgment can evaluate whether the company is still pointed the right way.
4. There are three types of decisions and they should not be treated the same way. Company decisions have long-term structural consequences and require deliberate slowdown. Tech decisions shape the platform the company will live inside for years. Feature decisions are generally reversible and should move fast. Treating all three with the same urgency is how organizations accumulate strategic debt.
5. Fix what’s broken, grow what’s proven, prove where to invest — and allocate capital accordingly. The three-bucket prioritization framework is not a ranking exercise. It is a capital allocation model. Knowing which bucket a request falls into tells you the appropriate investment level, the acceptable speed, and the right success metric before you have committed a single resource.
6. Shared framing of the problem is more important than the solution. A new CPO who rushes to reorganize the roadmap before understanding the decision rights of the company will rearrange the furniture without changing the room. The first 90 days should be spent mapping who owns which decisions and reaching agreement on how problems are described — not on what to build.
7. The greatest CPO alpha comes from the trajectory of the company, not from any individual product launch. A strong CPO ensures that good decisions compound and bad decisions do not become structural debt. The impact is not visible in any single quarter. It accumulates in the direction the company is heading two and three years out.
8. Cultural transformation is complete when everyone in the company starts speaking in outcomes, not outputs. The signal that product thinking has taken hold is not a new process or a new tool. It is the moment when people across functions — commercial, finance, member services — are instinctively asking about retention, conversion, and user impact rather than feature counts and delivery dates.
9. “Glues” are more powerful retention levers than hooks, and most organizations are not measuring them. When Nilesh looked past revenue and profitability metrics on a set of underperforming commercial products, he found they had outsized influence on retention. Reframing them as glues changed their priority entirely. The metric you optimize for determines what you see — and what you miss.
10. Cross-functional alignment is the product leader’s most underrated competitive advantage. Product has a vantage point that no other function has — visibility into every part of the business simultaneously. That position only creates value if the CPO uses it to reframe problems in ways that get the whole organization moving in the same direction, not just to arbitrate competing priorities.
11. Loyalty only works when it works for both sides of the marketplace. Over-optimizing for member satisfaction at the expense of merchant ROAS, or vice versa, creates a platform that wins one constituency and slowly loses the other. The intelligence layer Rakuten is building is designed to close that gap — matching incentives to intent in a way that creates value for members, merchants, and the platform simultaneously.
12. Easy money is the most dangerous kind of money to take. A partnership with guaranteed near-term revenue that covers a capability you need to develop in-house will delay that development indefinitely. Apple’s 20-billion-dollar search deal with Google left them without the skills they needed when search evolved into generative AI. The short-term yield always looks larger than the long-term cost — until it doesn’t.
13. The question is almost never whether an idea makes money. It’s whether building it is worth what you stop building. Bad ideas get filtered before they reach the CPO. What arrives at the top is a set of reasonable, potentially profitable options. The discipline is in evaluating not the individual merit of any one idea but the opportunity cost it imposes on the capabilities the company must own.
14. Transactions are moving off platform, and the cash back industry needs to be present at the point of payment wherever that happens. As AI and LLMs route more commerce outside of traditional apps and websites, the companies that have built trust and an intelligent commerce layer will have a right to play in those flows. The ones that did not invest now will be locked out of a transaction moment they can no longer access.
15. AI in a two-sided marketplace means targeted incentives replacing one-size-fits-all cash back. Programmatic loyalty is not a UX change. It is a fundamental shift in how value is allocated across members and merchants — moving from a static reward structure to a system that identifies the right incentive, for the right member, at the right moment, on the right surface.
16. Trust is not a feature. It is the product. For a cash back platform, the promise is getting paid. Every product decision at Rakuten is ultimately evaluated against whether it builds or erodes the confidence that members will receive what they were told they would receive. Owning that trust infrastructure — including an AI-based system to ensure order tracking regardless of cookies or browser behavior — is not a nice-to-have. It is the core.
17. Internal product teams have the same trap as external ones: becoming a service organization. When your customers are internal stakeholders, the pressure to fulfill requests accumulates just as relentlessly as with external ones. The discipline of product thinking — building a durable platform for future capability rather than solving only for the problem of today — is exactly as necessary inside an organization as outside of it.
18. The decision you’re really making when you over-index on the benefit is whether you trust your own hypothesis. Most requests arrive framed as “if we do X, we get Y,” and most people in the room focus on Y. Nilesh focuses on X — what actually has to change, why users would care, why merchants would pay for it, what is already proven. If X cannot be explained clearly, Y doesn’t matter.
19. The shift from transactional to intelligent commerce platform is the generational bet in the loyalty space. Cash back is becoming table stakes. The companies that win will be the ones that built the intelligence layer to match incentives to intent and the trust layer to guarantee that every transaction is honored — whether it happens inside or outside their own platform.
20. Product leadership matters most when it shapes what a company decides, not just what it builds. This is not a philosophical position — it is a structural one. The CPO sits at the intersection of every function, with visibility into every tradeoff, and the authority to say no to guaranteed revenue when it would hollow out a capability the company must own. That is the role. Everything else is coordination.