A product executive who builds. I've run the org, and I run my own AI systems every day.
20+ years leading product at consumer scale, most recently as Chief Product Officer and SVP at Yahoo!/AOL. Now I build AI systems and run product for teams: interim, fractional, or full-time for the right company.
From running product at Yahoo! to building my own AI system. One conversation.
Francois Richard hosts the @Scale podcast at Meta, where he also runs very large systems. He asked a Product Guy (me) on to talk about going from exec to builder, and what AI is doing to product.
We get into the products I shut down, the outage that stopped Yahoo! Mail, and what I told my own AI system on day one. Building got cheap. Judgment became scarce.
I come to it as a builder.
I spent 20-plus years building and running product at scale, most recently as Chief Product Officer and SVP at Yahoo!/AOL. Since 2022 I've held interim and fractional leadership roles, at EnterpriseDB, Globalization Partners and Owow.ai, seven to twelve months each. Since 2024 I've been building my own AI systems. For the right company, I'd take the role full-time. AI changed what a team can actually do, and most companies are still running the old playbook into a new environment.
In those roles the job is the one I did at Yahoo!: set direction, get product, design and engineering deciding together, and grow the people on the team. AI changes how a team operates. Someone still has to lead it.
I rebuilt my own operating model around AI first. Built the agents myself. Worked out how a system earns more access as it earns more trust. Read-only, then more, once it proves out. So when I sit with a leadership team, I'm drawing on things I run in production, not a deck.
There is one rule underneath all of it. The moment I notice I am the thing moving information between two systems, I stop and remove that step. I don't want to be the relay. Two years of doing that is how this got built, one irritation at a time.
I've been like this a long time. Right now it's a 1967 Fiat 500 in the garage. Work out how the thing actually runs, then make it run better. Twenty-plus years of running product orgs trained the same habit at a much bigger size, which is why I was ready when the tools showed up.
I was in a record store, saw an album I half-remembered wanting, and texted my system to ask if I already owned it. It checked 473 records and a wantlist and answered before I got to the counter.
That is the least impressive thing it does. It is also the one I tell people, because it shows the actual idea. Not a model that knows everything. A system that knows my things.
The same system runs the work. It knows my projects, my contacts, my commitments, and what I said about them six months ago. It sends me three things each morning instead of a list of forty, and it tells me when someone has gone quiet who should not have.
And it tells me no. It stops me most weeks, once for quoting a source I had not checked, once for a command that would have put a credential into a chat window. It has two years of my own mistakes in it, and that is what makes it worth anything.
I work with a small set of public and confidential clients. Some bring me in to lead product, interim or fractional. Some to rebuild how their team works around AI. Some for one decision they need to get right.
"An extremely capable operator who makes tough decisions."
Michael can lead multiple distinct product organizations, communicate clearly, and make the hard calls needed to move teams forward. He balances product excellence with strong business alignment.
"Strategic, decisive, and trusted at scale."
He led large, diverse product and design teams at massive scale while driving multi-billion-dollar performance. A rare executive who combines strategic clarity with operational discipline.
"Calm leadership through complexity."
Michael established structured operating discipline, led through reorganizations and ownership transitions, and drove impact with clarity and accountability.