The starter engagement
The first ninety days.
A quarter to find the work that just became feasible in your business — and ship some of it. Ninety days is long enough to map real workflows, put the first agents and AI workflows into production, learn from what shipped, and leave your team able to keep going.
Automating a task list doesn’t need a quarter. The ninety days are for expanding what your team can take on — with the people you already have.
Weeks 1–4 · Map and first ship
- Operating map of one business area, value stream, or initiative.
- A ranked roadmap of what just became worth doing — agents, AI workflows, long-deferred tools.
- The first workflow shipped or prototyped, in your environment.
Weeks 5–8 · Expand and embed
- Adjacent workflows brought into the same pattern.
- Measurement on what shipped, so improvement is visible.
- Vendor and tooling calls settled — including the ones to stop making.
Weeks 9–12 · Transfer
- Internal champions coached on the patterns, hands-on.
- A roadmap for the next quarter, current as of day ninety.
- A clean mutual decision about what, if anything, comes next.
There’s a checkpoint at the end of each month. If it’s not working, we stop — and you keep the map, the roadmap, and whatever shipped.
What you leave with
- A map of how the work actually gets done in the area we picked.
- Working agents and AI workflows running in your business, not in a slide deck.
- An internal champion who has seen the pattern and can carry it.
- Better judgment about where AI helps and where it creates chaos.
How we work together
- Weekly working sessions with leadership or operators.
- Async review of specs, workflows, and tooling decisions.
- Hands-on build alongside your team.
You don’t need to be an “AI person.”
This works the same whether the workflows live in engineering and product or in marketing, sales, and operations. What you need is a real business, real workflows that repeat, and someone inside the company who will work on this alongside us.
A useful comparison.
If you’ve seen how the AI platform companies embed forward-deployed engineers with their enterprise customers, this is that shape of work — with two differences. A vendor’s forward-deployed engineer works for the vendor; the job is to make the platform indispensable. We do the same embedded, ship-real-things work on your side of the table, with whatever tools actually fit.
The other difference is access: that model has typically only been available on enterprise-scale contracts. This is a version sized for founder-led companies — and the success condition is that your team keeps going without us.
Pricing.
Priced based on scope, fit, and shape of the work. We’ll talk through it on the first call and you’ll have a written proposal before anything starts.
How this fits with the rest.
The first ninety days is the usual way in — alongside ongoing AI operating partner work and initiative partnerships for single high-stakes bets. Same practice underneath.
See how engagements work