AI Clinic for Programmes and Portfolios
An accelerator, fellowship or fund books the clinic for its teams. Every team leaves with one number it measures, a simple test set and the date of its next decision.
Who it’s for: For accelerator and fellowship directors and VC platform teams who want their cohort or portfolio companies to measure whether their AI works.
The organiser pays. I don’t sell seats to individual founders or run open courses.
The cohort can build a prototype. It can’t tell whether it works.
Programme teams now build a prototype over a weekend, with coding assistants and off-the-shelf tools. They get stuck later: they don’t know what to measure, how to judge the quality of model output, or when to stop polishing and decide. In a survey before one of my webinars for a founder programme, 49 of 56 participants had already built an app with AI, and the thing they most wanted to learn was how to measure whether the product works.
What you get
- A group session: how to pick one thing to measure and how to judge the quality of model output
- 1:1 clinics for every team, on its own product and data
- For every team: the user job the product has to do, and the number that measures it
- A simple test set for checking quality after every change
- The date of the next decision: keep building, change direction, or stop
- A summary for the organiser: where the teams stand and who needs more support
How it runs
What the organiser provides
The clinic works if:
Price
from €2,800net · per edition
The starting price covers one edition: two sessions and 1:1 clinics for 8–15 teams, remote. The organiser is invoiced.
- a larger cohort becomes two editions, so every team gets its own clinic
- in Warsaw I can run the first session in person
Tell me how many teams are in the cohort or portfolio and when you want to start. Within 1–2 business days I’ll reply with a price and dates.
Or by email: michal@jaskol.ski
As of: November 2026
Before you write.
Is this AI tools training?
I don’t teach tools; teams usually use them already. The clinic covers what comes after the prototype: measurement, quality and deciding what to do next.
How many teams can take part?
Between 8 and 15 per edition. I split a larger cohort into two editions, so every team gets its own clinic.
Can it run online?
Yes, both parts work remotely. In Warsaw I can run the first session in person.
What does the fund get out of it?
A picture of the portfolio: which companies measure their AI results and which only show them, and where extra support pays off. Companies with a process ready for automation can move on to an Agentic Workflow Sprint.
What’s next
The organiser gets a summary of the edition. Teams with a process, an owner and data can move on to an Agentic Workflow Sprint or a Stalled AI Prototype Review.