Mateus Albuquerque
“Every 'quick custom field' request is a future migration you haven't scheduled yet.”
- Martin Knapic, Product MoatMateus Albuquerque has spent the back half of his career in enterprise B2B, the part of product where the roadmap has to survive contact with a sales team's promises. Twelve years in, he's unusually blunt about what enterprise buyers actually want from AI right now, which is mostly not what the demo suggested.
Identity & Background
- Name
- Mateus Albuquerque
- Current role
- Group Product Manager
- Current company
- Fernwave
- Location
- São Paulo, Brazil
- Years in product
- 12 years
Your story: how did you end up in product?
I was a sales engineer first, the person who got dragged into every deal the moment a prospect asked 'can it also do X.' I got good at saying no in a way that didn't kill the deal, and a VP eventually told me that was basically the product job. She wasn't wrong.
Core PM Philosophy
How do you define product success?
Net revenue retention, and specifically whether customers expand because they found a second use case on their own. Renewals tell me they tolerate us; organic expansion tells me they need us.
What's your decision-making philosophy?
I write the decision down before the meeting, including what would change my mind. It stops the loudest voice from winning and gives me something to review honestly six months later.
How do you manage stakeholders?
I make the cost of a custom request visible immediately, in the same room as the ask. Not 'we'll look into it' - a real number, tied to what it displaces. It changes the conversation from a favor to a trade-off.
Product & Decision-Making
How do you prioritize?
Enterprise buyers are asking for AI features they saw in a demo somewhere, without a real workflow behind the ask. Half my job is translating 'we want AI too' into an actual job-to-be-done before it goes anywhere near the backlog.
What's your favorite product you've shipped?
An approval-policy builder that let ops leads change routing rules without filing an IT ticket. Time to first custom workflow dropped from weeks to an afternoon, and support tickets about it fell to almost nothing.
AI & the Future
What's one thing about AI everyone gets wrong?
Vendors are still selling capability when buyers are actually asking about liability - who's accountable when the model's output causes a compliance problem. The vendors who start selling an answer to that question, not just a bigger model, are going to win the next wave of enterprise deals.
Company & Product
What does your product solve?
We're workflow and approvals software for mid-market operations teams - the layer that sits between 'someone requested something' and 'it's actually approved, budgeted, and tracked.' Our ICP is an operations or finance lead at a 500-5,000 employee company drowning in email-based approval chains.
What's the biggest misconception about it?
That enterprise customers want more features. Mostly they want fewer surprises - predictable behavior, clean audit trails, and a vendor that does what it said it would do next quarter.
What's the market opportunity?
We lose on breadth against the largest workflow suites, on purpose - we win by being implementable in weeks, not quarters, for a company that will never need that full surface area. Being narrower is the pitch, not the limitation.
More from the conversation
What do you do outside work that has nothing to do with software?
I restore old mechanical watches, badly, as a hobby. There's something calming about a system where every part has exactly one job and the failure mode is always mechanical, never ambiguous - the opposite of most days at work.
Any career detour you're glad you took?
Two years in implementation and customer success before I ever wrote a spec. Sitting in enough painful onboarding calls taught me more about what 'simple' actually means to a customer than any amount of design critique did later.
How big is the team behind it?
Product and design together are about 25 people, inside a company of just over 300. I run the enterprise segment specifically - three pods, each aligned to a vertical rather than a feature area.
How has AI changed enterprise buyer expectations?
Procurement teams now ask about model provenance and data handling before they ask about the feature itself. That question used to come from security three months into a pilot. Now it's slide two of the first call.
How is AI actually built into Fernwave's own product and internal process?
In the product, it drafts the first version of an approval policy from a plain-language description, which a human still has to publish. Internally, it's changed how we write specs for enterprise clients specifically - procurement now asks for a model card before they'll even schedule a security review, so that document exists before the feature does.
What's an underrated skill for enterprise PMs?
Reading a contract redline. The features your legal and security teams are quietly negotiating away tell you more about what enterprise customers actually value than any interview does.
Enjoyed this conversation?
A new interview with a product person lands every week. Get it in your inbox.