Skip to content

Focus area

AI & Product

Product management didn't get a new job title in the AI era — it got a higher bar. The PM has always stood at the intersection of five forces: the business, the technology, the delivery process, the customer, and internal stakeholders. AI walked into every one of those rooms and raised the stakes: pricing is simulated against live data, roadmaps must account for model drift and inference cost, "done" is now continuous evaluation, customers increasingly meet the product through an agent, and compliance needs ongoing visibility because behavior can change after release. This pillar is what I've learned building AI products in one of the hardest environments there is — regulated insurance — distilled into things you can use.

01

Traditional PM vs AI PM

A side-by-side of how each of the five forces shifts with AI — business (market literacy → model literacy), technology (deterministic → probabilistic trade-offs), delivery (one-time acceptance → continuous eval/monitoring), customer (periodic research → live signal streams and agent interfaces), stakeholders (sign-off at launch → continuous compliance visibility).

The conclusion: this isn't a specialization that spins off into "AI PM"; it's the new baseline for good product management.

Proof

  • Flagship essay: Stop Calling Them "AI Product Managers" (Insights).
productaicareers
Read the flagship essay

02

Building AI-era PM Skills

The capabilities to layer on top of classic PM: model and data literacy; evaluation and monitoring discipline (golden sets, red-teaming, drift); designing human-in-the-loop controls; reasoning about cost, latency, and explainability; and regulatory judgment.

Who has the shortest gap: business-savvy PMs (add AI/data literacy), then business stakeholders (add engineering/data fluency), then technical stakeholders (add business ownership — pricing, regulatory judgment, commercial trade-offs).

productaiskills

03

The Zero-to-One AI Product Playbook

My framework for taking an AI concept to a shipped, adopted, monetized product: start from the problem and the P&L, not the model; ship a thin vertical slice on a real domain schema; put evaluation and governance in from day one; keep humans on judgment and money; and design pricing/GTM alongside the product, not after.

Grounded in building $1M ARR in 18 months and incubating a $6M-funded startup.

Proof

  • Built $1M ARR in 18 months (Swiss Re).
  • Incubated a $6M-funded startup, 4→25 team, 6 beta → 3 paying (Cognizant Property Insights).

04

For Hiring Managers — Finding the Right AI-era PM

What the evolved PM (the "Product Navigator") actually looks like, and how to assess them: signals of model literacy without hand-waving, evidence they've shipped and measured, comfort with probabilistic trade-offs, and the judgment to defend a decision to a regulator or a room of stakeholders.

A practical guide for leaders building AI product teams.

producthiring

Signature frameworks

Original models, from real work

Models → Processes → Outcomes

My operating lens: technology only matters when it reshapes a process and lands a measurable business outcome.

When to use it: Framing any AI-in-insurance initiative end-to-end.

Zero-to-One AI Product Playbook

Idea → shipped, adopted, monetized AI product — problem and P&L first, evals and governance from day one.

When to use it: Launching a new AI product.

The Product Navigator

The AI-era PM archetype: business-grounded, AI-instrumented, still in command of the five forces.

When to use it: Hiring or growing AI product talent.