Adam Fishman on Building Innovation Labs That Actually Kill Things

The Innovation Executioner

Adam Fishman arrived at Mozilla with a mandate most innovation leaders would find uncomfortable: figure out which projects deserve to die. The company had accumulated a portfolio of technically impressive experiments with no business model, no go-to-market strategy, and no clear reason for existing beyond "cool technology." Some had dozens of engineers working on them for years.

Fishman's assessment process was brutal in its simplicity. "When I came into the company, one of my jobs was to help us look at things retrospectively, retroactively, and go, is there anything here? Who are we selling this to? How will we make money from this? What do we think the long-term perspective on this market area is?" The answer, most of the time, was silence. "A lot of times we didn't have answers to that question, and so it meant shut it down."

I was a little bit of the killer of projects for a while. But honestly, you kind of have to get back to square one and rebuild and start working through this process again.

— Adam Fishman

This wasn't the work of Innovation Kevin, the corporate archetype who makes beautiful slides about the future and never ships anything. Fishman came from Patreon, where he helped scale the company to a billion-dollar valuation, and Lyft, where he built city expansion playbooks. At Mozilla, his portfolio includes everything that isn't Firefox: an AI review-scoring product called Fakespot acquired a few years ago, a builders program funding local and open-source AI companies, and multiple innovation studio pods iterating on zero-to-one products.

Stage-Gating as Venture Discipline

The problem Mozilla faced wasn't a shortage of ideas. It was a complete absence of filtering criteria. Teams would secure funding based on pitch quality or timing rather than evidence. Technical feasibility trumped market viability. Projects ballooned to 50 or 75 people without ever answering basic questions about monetization or customer acquisition.

Fishman introduced a five-stage framework that mirrors venture capital stages: discovery, exploration, viability, growth, and sustaining. Each stage has specific hypotheses to validate and unlocks new funding tranches. A team evaluating body—what Fishman calls "the venture board"—decides whether projects advance or die.

Discovery and exploration are coupled. Discovery defines a problem worth solving and validates market attractiveness. Exploration prototypes solutions and confirms technical feasibility. Teams might cycle here for two to six months, but they're moving fast, discarding ideas, iterating. "AI makes it really easy to prototype and test stuff super fast," Fishman notes. The bar is simple: prove Mozilla has a unique reason to play in this space.

I would rather prematurely shut something down than flail away at something that really didn't have any long-term vision for 12 months without any end in sight. And it actually doesn't help anyone to do that.

— Adam Fishman

Viability is where teams validate product-market fit, iterate on value propositions, and prove the product can acquire and retain customers. This can take a year or two. Growth means scaling existing loops. Sustaining means optimizing for profitability and EBITDA. The entire structure is designed to prevent what Fishman calls "innovation theater"—work that looks like progress but generates no revenue or strategic clarity.

Killing Dreams Requires Executive Cover

Fishman is upfront about the emotional toll of this work. Product managers operating in this system function like startup founders facing a skeptical board every few months. "You're basically pitching your idea and regularly coming back in front of kind of an advisory board who's saying, we're not there, pivot, shut it down," he explains. "You're getting constant barrage of like, okay, now what?"

The role demands executive support. Fishman's CEO at Mozilla didn't see it as her job to decide whether a product should live or die. That authority sat with Fishman and his advisory board. Without that cover, the framework collapses into politics—whoever shouts loudest or presents best wins the budget.

The shift also required financial rigor. In the past, teams would request funding for engineers or infrastructure without clear success criteria. Fishman worked with Mozilla's CFO to define what has to be true at each stage to unlock the next funding round. The amounts aren't Silicon Valley scale—"no one's throwing $100 million at a random AI tool," he jokes—but the discipline mirrors venture-backed startups. Teams that can't answer basic market or customer questions don't advance.

It's hard to be a PM in this world. If you think of the PM as sort of the founder of a new product or company or whatever, you have to be pretty resilient.

— Adam Fishman

The Fakespot Acquisition and the Builders Program

Not everything in Fishman's portfolio came from internal ideation. Fakespot, an AI-powered review-scoring tool, joined Mozilla through acquisition. The product analyzes Amazon reviews and assigns trust scores, helping users distinguish legitimate feedback from fake noise. Since the acquisition, the team has expanded into broader AI safety and security work, including in-browser agent experiments.

The Builders Program represents a different bet: funding external teams working at the intersection of local and on-device AI. Mozilla provided non-dilutive funding to 14 or 15 companies, culminating in a demo day. Fishman sees it as a way to stay close to cutting-edge work that can't happen inside corporate walls. Some companies may become acquisition targets. Others will develop technology Mozilla can adopt or invest in further.

Both initiatives reflect Fishman's broader thesis: Mozilla can't rely on Firefox forever. The browser remains strong, but the company needs second and third acts to future-proof itself. "Who knows what could happen to the browser in 10 years from now? Like maybe we don't use web browsers anymore," Fishman says. Building new revenue streams that can self-fund is the long-term goal.

Jovial, Not Soft

People who meet Fishman often misjudge him. He laughs frequently, keeps things collegial, doesn't try to intimidate. That demeanor leads some to assume he's not serious about winning. "That could not be farther from the truth," he says. "I have very high expectations for the work that people get done when they work with me, on a team with me, for me, like whatever."

The tension between approachability and intensity defines his leadership style. He'll kill your project without drama, explain why in plain language, and move on. He won't let teams spin their wheels in research purgatory, chasing unattainable certainty. He won't fund projects because they're technically interesting or because someone made a compelling deck.

We can have a fun kind of collegial relationship. And at the same time, I have very high expectations for the work that people get done.

— Adam Fishman

Fishman's interim role at Mozilla is wrapping up. A new SVP named Peter is taking over. But the stage-gating framework remains—a forcing function that treats innovation as a portfolio of bets, not a laboratory for cool ideas. In an industry flooded with innovation theater, Fishman built a system that defaults to killing things. That might be the most innovative part of all.