How we work

Product Teardown

$2,5005 business daysOne artifact per day.

Five days, five artifacts, one decision. By day five you know the single thing worth building next and you have a prototype or a clickable spec for it. Credited in full against a Build Sprint.

The full $2,500 is credited against a Build Sprint if you decide to build with me.

Day 1
Understand the product the way an operator does. I walk the real flow, not the pitch deck: signup, first value, the paid path, and where the AI actually shows up.
Artifact
A one-page product map: the core loop, who it is for, and the moment a new user either gets value or leaves.
Day 2
Find where the product leaks. I look at activation, retention, pricing, and the AI surface, and mark the drop-offs that cost the most.
Artifact
A leak list across activation, retention, pricing, and AI surface, ranked by revenue at risk.
Day 3
Decide the one thing worth building next, and say plainly why the rest can wait. One bet beats a backlog.
Artifact
The one-thing brief: the single build to run next, the tradeoff behind it, and the KPI it should move.
Day 4
Make the bet real. Depending on the build, that is a working prototype on your own data or a clickable spec a team can pick up on Monday.
Artifact
A prototype on your data, or a clickable spec with the screens, states, and edge cases drawn out.
Day 5
Put it all on the table and defend it live. You leave knowing exactly what to build and what to ignore.
Artifact
A 30-minute readout: the leaks, the one thing to build, the prototype or spec, and the numbers behind the call.

Out of scope

No production build. That is the Build Sprint.
No ongoing calls or Slack access after the readout.

Build Sprint

$15k to $40k6 weeksOne artifact per week, a KPI agreed in week 0 before signature.

Fixed scope. Priced by the sprint, not the hour.

Six weeks, one system, one KPI. We agree the number before you sign. Each week produces one artifact, and week six ships it to production with a handover doc your team owns.

Week 0 KPI

Week 0 is a KPI conversation. We name the one metric this build has to move and agree it before signature. No number, no sprint.

Payment

40 / 40 / 20 against milestones.

A feature or agent that lives inside your product, built on your stack, shipped to production and handed over. This is how I built Surge's RSVP, check-in, and membership flow: real users on day one, not a demo.

Week 0
Before you sign, we agree the one KPI this surface has to move and how we will read it. If we cannot name a number, we do not start.
Artifact
A signed one-KPI scope: the metric, the baseline, the target, and what is explicitly out of scope.
Week 1
Set the shape. The user flow, the data model, the model and vendor choice, and how the AI output gets checked before a user sees it.
Artifact
A build spec plus the schema and a wired data path, running against your real data.
Week 2
Build the core loop end to end. The input, the model call, the output, and the guardrails around it.
Artifact
A working core loop behind a flag, doing the real thing on real inputs.
Week 3
Make it hold up. Edge cases, empty states, bad inputs, and the failure paths a real user will hit.
Artifact
The feature handling the messy cases, with evals that catch a regression before a user does.
Week 4
Put it in front of real usage behind the flag. Watch what actually happens against the KPI baseline.
Artifact
A staged release with a live read on the KPI and a short list of what to tighten.
Week 5
Tighten on evidence, not opinion. Fix what the usage data flagged and cut what nobody touched.
Artifact
The tuned feature plus a dashboard tracking the KPI you agreed in week 0.
Week 6
Ship it to production and hand it over so your team owns it without me in the loop.
Artifact
Live in production, plus a handover doc: architecture, the model and vendor calls, the evals, and how to run it.

Out of scope

No open-ended scope. One system, one KPI, fixed price.
No staff augmentation. I build the system and hand it over.

Retainers

Two ongoing options. Each has a weekly, monthly, and quarterly rhythm, so you always know what lands and when.

Fractional Product Lead

$8,000 per month3-month minimum, about 20 hours a week.

I run product for you part time: roadmap, specs, prioritization, AI architecture, and hands-on build when that is faster. You get a product leader, not a deck.

Weekly

A prioritization pass: what your engineers build this week and what waits.
Specs written ahead of the work, so nobody is blocked on a decision.
Hands-on build when shipping it myself is faster than explaining it.
A named escalation path, so you always know who to ping and when.

Monthly

A roadmap review against the numbers, and a re-cut where the data says so.
The model and vendor calls: what to keep, what to swap, what it costs.
A written update on what shipped, what moved, and what is next.

Quarterly

A step back on product strategy and where the next bet should land.
An honest read on whether a full-time hire now beats the retainer.

Out of scope

No hourly work and no staff augmentation.
No SMB automation subscriptions.

Product Care

$1,500 per monthAfter a build. Cancel with 30 days.

Keep what we shipped healthy: monitoring, evals, regression checks, and model and prompt upkeep, with a monthly KPI report. The quiet retainer that keeps an AI product from drifting.

Weekly

Monitoring on the live system, so drift and errors surface before a user reports them.
Evals run against the golden set to catch a regression early.

Monthly

Model and prompt upkeep as vendors change under you.
Regression checks on the paths that matter most.
A KPI report: what the system did this month and what it cost.

Quarterly

A health review of the whole surface: where it is aging and what to shore up.
A short list of small fixes worth doing before they become big ones.

Out of scope

No new features. New scope is a Build Sprint.
No roadmap ownership. That is the Fractional Product Lead.