Samantha Nair
“The org chart stopped being the interesting design problem. The human-AI handoff is.”
- Martin Knapic, Product MoatSamantha Nair has spent sixteen years in product, the last four of them rethinking what a leadership job even is once half the org's decisions run through a model before they reach a person. I wanted to open the series with her because she's unusually honest about the parts of leadership that got harder, not just the parts that got automated away. We talked about the move to Barcelona, the instinct she had to unlearn, and what she's actually looking for when she hires now.
Identity & Background
- Name
- Samantha Nair
- Current role
- VP of Product
- Current company
- Solventra
- Location
- Barcelona, Spain
- Years in product
- 16 years
Your story: how did you end up in product?
I started in ops at a logistics company, the person everyone came to when a process broke. A director noticed I kept redesigning the workaround instead of just running it, and asked if I'd ever thought about product. I hadn't - I didn't know it was a job title yet. Six months later I had one.
Core PM Philosophy
How do you define product success?
Whether a treasurer trusts our numbers enough to act on them without checking a second source. Revenue and usage follow from that, but trust-to-act is the one metric I'd keep if I could only keep one.
What's your decision-making philosophy?
I separate decisions by how reversible they are. Reversible ones get made fast by whoever's closest to the problem; irreversible ones - anything that moves money or changes a customer's data model - get a written pre-mortem first.
What's your biggest PM lesson or mistake?
Rewarding certainty. I used to promote the PM who walked into a room with the most confident answer. Now I'm more interested in the one who can tell me exactly how confident the model's answer is and why, because that's the skill that actually prevents expensive mistakes.
How do you manage stakeholders?
I stopped treating stakeholders as an audience and started treating them as inputs with a deadline. Finance, legal, and sales each get a standing 20-minute slot before anything reaches roadmap review, so disagreements surface early and small.
Product & Decision-Making
How do you prioritize?
Customer outcome first, then reversibility, then cost. If two bets both move an outcome, I pick the one we can undo. Anything with a model in the loop also has to show a plausible eval story before it can rank.
How do you do customer discovery and validation?
A treasurer will tell you what they wish for in an interview and do something different in the tool. So we sit in on real month-end closes and watch what they open in a second tab - that second tab is the roadmap.
What's your favorite product you've shipped?
The cash-position dashboard that replaced a seven-tab spreadsheet for our first ten customers. Nothing about it is clever; it just removed a Monday-morning ritual that everyone hated and nobody had questioned.
AI & the Future
How is AI changing product management?
Less about writing the spec, more about defining what 'good' looks like and who signs off on it. PMs are becoming editors of judgment - the job has more taste and less transcription than it used to.
Which AI product do you use and respect?
Claude for drafting and pressure-testing memos - I'll argue against my own first draft with it before anyone else sees it. And whichever coding assistant my engineers are raving about that month; I like understanding what they're actually seeing.
What's one thing about AI everyone gets wrong?
That the hard part is capability. It isn't - capability is a commodity by now. The hard part is deciding where a wrong answer is cheap and where it's catastrophic, and designing the handoff accordingly.
Beliefs & Opinions
What's your unpopular product opinion?
Accountability when a model is in the decision loop. We've gotten good at measuring model accuracy and terrible at defining who owns the outcome when the model was right, the human overrode it, and the override was the mistake. That's a leadership and process question, not a technical one, and most orgs don't have an answer yet.
Which author, thinker, or influence shapes you?
I stopped trying to track every model release and instead track three things weekly: what my best engineers are excited about, what our support queue is telling us, and what one trusted peer group is seeing. Filtered signal beats raw volume every time.
What makes a great product?
It removes a decision the customer never wanted to make. The best products I've seen feel less like a tool and more like something that's already been handled.
Company & Product
What does your product solve?
Treasury and cash-management software for mid-market companies - the tools finance teams use to see where their money sits across banks and currencies, and move it without six spreadsheets and a prayer. Unglamorous, high-stakes, and surprisingly under-served.
What's the biggest misconception about it?
That finance software is boring by nature. It's boring by neglect. The people using it carry real stress and real personal risk, and there's a lot of room for products that respect that.
What's the market opportunity?
We're not trying to be a treasury module built for a company ten times our customer's size - we're for the mid-market company that's outgrown a spreadsheet but would be a rounding error to an enterprise vendor's sales team. Our ICP is a 200-2,000 employee company with a treasurer wearing three other hats. Being the right size for them, not the biggest platform in the category, is the whole positioning.
Why would you recommend it to a peer?
If you're in a mid-market finance or treasury role and spend the first week of every month reconciling, it's worth an hour. Even if you don't buy, you'll leave with a sharper sense of what your process shouldn't require.
More from the conversation
What's something outside work that quietly shapes how you lead?
Sailing, badly. I crew a small boat most weekends and there is nothing more honest than wind - it does not care about your quarterly plan. It's taught me more about reading conditions and adjusting in real time than any leadership book has.
Looking back, what almost pulled you away from product entirely?
A stretch running a P&L directly, no product title at all. I liked the clarity of owning a number, and it's why I still think every product leader should carry one at some point. I came back to product because I missed building the thing, not just measuring it.
How big is the org you're leading product for?
About 40 people across product, design, and research, inside a company of just under 600. Four product pods, each owns a slice of the treasury workflow end to end rather than a horizontal layer, which is deliberate - it keeps teams close to a real customer outcome.
Where has AI actually landed in the product itself, not just the roadmap slide?
Cash-flow forecasting and anomaly detection on transactions - both places where a model can propose and a human treasurer has to approve before money moves. Nowhere in the product does an agent move funds unsupervised, and that line is not going to move soon.
How has it changed the way your own team works day to day?
Half our internal spec review is now spent agreeing on eval criteria instead of pixel details. PMs pair with a data scientist earlier than they used to, sometimes before there's a mock at all, because the model behavior has to be scoped before the screen does.
Where do you think this is heading over the next few years?
Toward much smaller, more specialized product teams doing what used to take an org twice the size, with the leadership skill shifting from headcount planning to judgment calibration - knowing when to trust the model's output, and teaching your team to know it too. The org chart gets flatter; the trust chart gets more important.
What do you look for when hiring product leaders now?
Whether they've ever had to kill a feature that tested well but that they didn't trust. AI makes it cheap to generate plausible-looking wins. The judgment to say no to one anyway is the same judgment it's always been - it's just being tested more often now.
Enjoyed this conversation?
A new interview with a product person lands every week. Get it in your inbox.