Working together
Start with the problem
you can already measure.
The journal is where I think the work through. I also do that work with other teams: on agents, AI systems, release gates and the engineering decisions around them. This page is the way in.
Evidence first.
Then the build.
Most engagements start the same way. We pick a number you already track, like cost per task, time per invoice or error rate. I measure where it stands from your own records before anything is built. Every finding points to the record behind it, and anything I can’t confirm is listed as unconfirmed.
From there the work takes one of three shapes: a defined project with a clear finish, fractional leadership as a part-time CTO or AI lead, or advisory for a founder or leadership team. Consulting engagements run through my practice at transition2.ai ↗, where each first step is described in full.
I don’t take a referral fee or reseller margin on any model or tool I recommend.
Pick the one
closest to your week.
Each link opens a short tool on transition2.ai. It gives you a first read on your situation and points to a next step if you want one.
Your agent works in the demo. It has to hold up in production.
A fixed-price review of an AI product before a launch or a raise: findings ranked by severity, each with the evidence behind it and a plan to fix it.
Most of your app was built with AI, and launch is close.
Check where an AI-built codebase is likely to break before real users find it.
You need to know whether your agent is ready for real users.
Score where your agent stands before you put it in front of customers.
Your tech lead left, or you never had one.
See what the first month of interim CTO work looks like, as a sample plan.
The LLM bill is growing faster than usage.
Work out what one task actually costs you today, so you know which number to bring down.
An AI pilot started well and then stalled.
Work out why it stalled and what would have to be true to restart it.
Not here? There are tools for finance teams, clinics, trades, agencies, online stores and investors too. See all the tools ↗
A 30-minute call,
to see if it fits.
We talk about what you’re building, the number or decision that matters, and whether a first step makes sense. The booking page is set up for AI audits, but it works for any of the options here. Use its questions to tell me what you have in mind.
A few lines
is enough.
Tell me what you’re building, where it stands, and what you want help with. It goes to my consulting practice at transition2.ai, and I’ll reply by email.
Thanks, it’s sent.
Your note reached transition2.ai. I’ll reply to the email address you gave. If you’d rather pick a time now, you can book a call ↗.
Things I’m building
alongside the writing.
These are products I work on. Each has its own site with its current status. They are separate from my consulting work.
FrameBright
Digital signage for Android. One locked, remotely managed player for PowerPoint (with supported interactive links), video, web pages, images and HTML, set up with a guided desktop app.
In public preview.
GoodQA
Guided human QA. It scans a GitHub repo, generates step-by-step test missions, records the tester, and turns a failed step into a bug report a developer can act on. Agents can use it through a REST API and MCP.
Live. Paid plans aren’t open yet.
OptinRouter
An opt-in exchange for AI coding work. Developers choose what gets captured, and approved buyers license the cleaned work. Tracking, permission to sell and accepting a benefit are separate choices.
Private beta, invite only.
Petfiliate
A pet-only creator affiliate network from the team behind P51 Media. A person reviews every application. Approved creators get tracked links to Chewy, the first live retailer, and follow clicks, orders and commission status in one place.
I’m the CTO. New creator applications join a waitlist.