7-day build · $5K setup + $1,497/mo · Done for you

Build your first Play in 7 days

I build your Employee Zero and ship your first Play on your real product. Scoped on Day 1. Running by Day 3. Yours by Day 7. Tuned by me every month after.

Anatomy

How a Play actually works.

Employee Zero is the agent. It holds your context. Product, customers, positioning, voice. A Play is a workflow the agent runs. Scoped to one PMM job. Producing one specific artifact. Always on.

Inputs

Your context

loaded on Day 1

  • Product

    What you sell, plainly

  • ICP

    Who it's for, and who it isn't

  • Positioning

    Category, alternatives, wedge

  • Messaging

    POV, VP, benefits, proof

  • Voice

    Words you use, words you never use

  • Proof

    Quotes, metrics, customer names

lives in CLAUDE.md

The agent

Your Employee Zero

holds all the context

The brain that never sleeps. Never asks you to re-explain your product. Already sounds like you.

Runs your Play

INPUT spec drop / cron / competitor move
PROCESS pull context → generate → score → rank
OUTPUT artifact committed + Slack notify
Output

Artifacts you can demo

starting Day 3

  • Launch brief, running on your product

  • Tested message variants, ranked by ICP fit

  • Battle cards rebuilt from sources you approve

  • Messaging drift flagged across your channels weekly

ships to Slack · GitHub · Notion

Pick your first

On the kickoff call, you pick one of the three Plays. Each one is scoped, spec'd, and battle-tested. Build the next ones yourself. The foundation's already there.

Compete

INPUT A competitor ships a feature, a pricing change, or a release note
PROCESS Pulls the real sources, rebuilds your positioning against them, and you approve before it ships
OUTPUT A fresh battle card, objection handling, and landmine questions for sales

Position

INPUT Your live website, ads, emails, and sales decks, checked every week
PROCESS Scores every asset against your locked positioning and flags anything that has drifted
OUTPUT A weekly drift report in Slack with the exact lines to fix

Launch

INPUT A product or feature you are getting ready to launch
PROCESS Writes the brief and positioning for your ICP, then maps the timeline to make the launch land
OUTPUT A launch brief, the positioning, and a dated timeline you can run
The 7 days

What happens, day by day.

01 Day 1

Kickoff + Context Load

45-minute call. We define the Play scope. I load your CLAUDE.md with product, ICP, positioning, voice, messaging hierarchy. Whatever you have today, we use.

02 Days 2–4

Agent Build

I build your Employee Zero and scope your first Play. You get async updates on Slack. By Day 3, the Play is running on your real product. Not a toy example.

03 Days 5–6

Tune + Trigger

We review the first live outputs together. I wire triggers and integrations (Slack, GitHub, Notion, cron). You start using it.

04 Day 7

Handoff

30-minute handoff call. You own the code, the CLAUDE.md, and the Play. You walk out running it, not waiting for me.

Week 1 deliverables

A working system you can demo.

Custom CLAUDE.md built with you (your product, customers, ICP, positioning, voice, proof)

Your Employee Zero (the agent that holds all that context, forever)

One Play scoped, built, and tested on your actual product

Triggers + integrations wired (Slack, GitHub, Notion, cron, webhook)

System docs you own. The repo lives in your GitHub from Day 1

Priority access to new Plays (2 weeks before public release)

Service guarantee. I keep working until you can demo it.

After Day 7 · Every month

The Sprint ships it. The retainer keeps it sharp.

Most agents go stale within 30 days because nobody re-trains them. Your product evolves, the context drifts, and the output gets generic again. The retainer makes that my job.

01

Tuning + maintenance

When output drifts, I retune. New product launch, new ICP, new positioning, I update the context for you. You don't have to remember to.

02

Slack on tap

Async by default. Drop a question, get an answer. No daily check-ins, no fake urgency. Real PMM problems, real fixes.

03

Monthly review call

30 minutes a month. We check the Plays, surface what's not working, and scope opportunities for new Plays.

04

Drift detection + context refresh

Your product evolves. So does the agent. I refresh the CLAUDE.md when it matters and re-test the Plays against the new context.

Want a second Play? Book a Training session ($2K, 2 × 2-hour blocks) where we scope and ship Play #2 live, or scope it together on the next monthly review call. Once Employee Zero exists, every new Play is a fraction of the effort.

The Sprint

$5,000 setup · 7-day build
+ $1,497/month ongoing retainer

Four Sprint spots open per month. When they're gone, the waitlist reopens the following month.

Not ready for a Sprint? Install a free skill first →

Newsletter

The builds. The systems. The results.

What I'm building, what's working, what flopped. No fluff. Just the work.

No spam. Unsubscribe anytime.

Sprint FAQ

The deeper questions.

Q How many Sprint spots open per month? +
Four. When they're full, the waitlist reopens the following month.
Q Do I need to be technical? +
No. If you can run a Slack message and read a markdown file, you can run Employee Zero. I handle the build. You handle the strategy and the feedback.
Q Can I add more Plays after the first one? +
Yes. Book a Training session ($2K, 2 × 2-hour blocks) where we scope and ship a second Play live, or scope additional Plays together on a monthly review call. Once Employee Zero exists, every new Play is a fraction of the effort.
Q What if my product changes mid-month? +
Tell me in Slack. I refresh the CLAUDE.md context, retune the Plays, and re-test the output. That's exactly what the retainer covers.
Q Can I pause the retainer? +
Yes. Pause for a quarter, resume when your product needs the agent again. The system stays yours either way. You own the code from Day 7. The retainer only keeps it sharp.
Q What's the service guarantee? +
If you can't demo a working Play by Day 7, I keep building. The retainer doesn't start until your system is live and producing output you'd actually use.

Question not here? Book a call and ask it live.