From the first call to a working pilot.

What happens after you book a call: each step, what you get from it and when. The timings are from a real AI project in November 2026; the client's details are left out.

  • 3 daysRequest to call
  • 58 minDiscovery call
  • Same eveningDiscovery document
  • 3 weeksPilot plan

Day 0
Request.

You book a call or send a message. The form asks for your company, your role, the timeline and what the call is about, so I know before we talk whether this is a role, a project or a question I can answer by email.

I read every request myself and confirm it, usually within a day.

You get: a confirmed time, or a straight answer if I am not the right person for it.

Before
the call.

I go through your website, product and public pages and write a one-page brief for myself: who you are, what you probably need, the questions for the first ten minutes, and the things I will not promise until I have tested them.

You get: a call that starts on your problem, not on a generic pitch.

The
call.

You do most of the talking. I ask how your customers describe the problem in their own words, what data exists, who has to approve what, and what a good result looks like in numbers.

With your permission I record and transcribe the call, so I can quote you exactly afterwards instead of working from memory.

You get: a clear yes, no or "not yet" on whether I can help, before you spend anything.

Same
evening.

A written discovery document goes to your inbox. It repeats the problem in your words, shows how the solution would work, and puts a few product directions on your own pages so you can see them rather than imagine them.

Sections of the real one, in order:

  • How it fits your product and the way you already sell
  • Your customers speak in sentences, with examples from the call
  • How it works, in plain steps
  • Product directions, each drawn on your real pages
  • What else the same data enables
  • Where the market is going and why design decides it
  • Speed: what the customer waits for
  • A 3-week pilot and the open questions

Research
and tests.

Before writing a proposal I check what I would be promising. Whether the work can start before you send me any files. Whether the hard cases you mentioned in the call actually work, such as a size selling out at noon. Which AI models answer your own sample questions best for what they cost.

You get: a proposal built on things I have tested, not on assumptions.

The
proposal.

It opens with what I found and how the examples you gave are solved. The plan and the fee come after that, so you can judge the fee against the work instead of the other way round.

  • What I found on your site and why things fail today
  • What I build: what your customer sees and what your team sees
  • The hard cases from the call, answered one by one
  • Week-by-week plan and how success is measured
  • Model choice by price and performance, compared on your data
  • Infrastructure: where it runs and who pays the running costs
  • Fee and payment, and what happens after the pilot
  • Questions your own clients may ask, with answers

Pilot and
handover.

The pilot runs in short weekly steps, and you see progress on a working link, not in status reports. You try a working demo before the final payment. Once it is paid, the code, the hosting and the documentation move to you.

You get: a system you own and can run without me, or keep improving with me.

Start
here.

Tell me about the problem on a 30-minute call, or in a message if you prefer writing.

Book a callSend a message

Have a problem worth solving?Let's make it reliable

Start a conversation

Open to full-stack and applied AI roles, remote or hybrid.

Tell me about the role, the team and the problem.