Bayesian Leaf

Overview

Hew Suber, MBA

Fractional product and design leadership

I set product direction for teams building with AI, and stay close enough to execution to back the call.

Seventeen years across Meta, Pinterest, Opendoor, and early-stage companies. Every engagement is with me. I hold the direction and move into the work without handing off the context, so the team can ship without me.

On a recent engagement, the prototype went to engineering as the spec, and a scoped project became standing product and design direction.

What it sounds like when someone needs me

The demo works. Real customers still need us on the call to get value. We're hiring engineers, but nobody actually owns product. We can't get it to self-serve. Every account needs a person in the loop. The model performs, but people don't trust the output enough to act on it. The roadmap is full and nobody can say which thing carries the outcome.

Who they are

Stage

Seed to Series B. Roughly 10 to 100 people.

Kind of company

Building with AI. Products that decide, recommend, or automate. B2B software moving toward self-serve.

Who I talk to

Founder, CEO, or CTO. Sometimes the lead product person, acting as sponsor.

What I actually do

  1. Map the system. Most of the problem is already visible. It has rarely been drawn in one place.
  2. Find the load-bearing decision. The product is usually waiting on one decision. I find the one carrying the weight.
  3. Make it legible. A person should know what to do, and why.
  4. Build the trust layer. Evidence, controls, and a way back when it's wrong.
  5. Ground it in measurement. If we can't tell what changed after release, we don't know whether the decision helped.
  6. Sequence it. Each release makes the next one easier.

Engagements

Direction Sprint
Four weeks, fixed scope. $30K. A direction document, a high-fidelity prototype with the direction inside it, and the sequence the team can carry.
Embedded
Hold the direction while the team ships. Full scope from $20K/mo. Advisory through hands-on execution, with the monthly scope setting priorities, ownership, and price. The context does not reset.
System Read
One week, fixed scope. $12K. A written read of the codebase, the decision the product is waiting on, and a high-fidelity prototype that proves the smallest important point.

When the scope includes direct product work, I begin with a read-only pass through the codebase, so the work starts from what is already true.

How engagements start

A 30-minute working session. No deck, no pitch. Bring whatever is stuck, and we find the one decision the rest is waiting on. I scope the proposal after I understand the work.

Book a working session Download the one-pager (PDF)

Pass this along

If you know a team this fits, send them the link. Nothing here asks them for anything first.

Contact

hew@bayesianleaf.co bayesianleaf.co Atlanta, GA, working worldwide
Book a working session
Download the one-pager (PDF)

Link copied.