About
Most school leaders learn the job on real students and staff, and many say they started unprepared. Medicine and aviation have long used simulation to train for the moments that matter. We are bringing that kind of practice to school leadership, with an honest account of what the evidence shows.
Principles
A design that breaks one of these is reworked before it is built.
01
No model call changes the school's true state. Agents propose, and rules resolve.
02
No real person's data goes in, and nothing comes out that ranks or selects a school leader.
03
Any outcome leads back to the rules, random draws and decisions behind it.
04
Every school system runs on the same engine, and each pack holds what differs.
05
Each build gate rests on published figures, leaders' panels and the build's own tests, and its results are published. The pilot trial is published whatever it shows.
06
Serving school leaders review every pack, and panels of them judge whether the school is recognisable.
Who is building it
Synthetic School is built by one founder working with AI coding agents. Each stage of the build ends in something a buyer can use, so the plan can pause at any gate without waste.
Founder
Jamie Scobie is a secondary school senior leader who leads on attainment and inclusion and builds the data platforms his colleagues use every day. He follows the research and policy debate on AI in education and writes about it. He holds two master's degrees with distinction, one of them from the University of Cambridge.
Get involved
The build gates depend on people from outside it. Each is asked for a bounded piece of time.
| Who | What they do | Time asked |
|---|---|---|
| A serving school leader from each pack's system | Reads the pack and its persona library | A few hours, agreed in advance |
| A child protection lead for each pack | Reviews the storyline library, then once a year | Agreed per pack |
| Panel school leaders | Play scenes, judge blind mixes of agent and human transcripts, and predict outcomes | About 2 hours a round |
| Teachers | Play the same scenes as the agents, to make the human transcripts for the blind mix | Agreed per round |
| An academic partner | Co-designs and publishes the validation studies, and runs the pilot trial's evaluation independently | Agreed with the partner |
| Design-partner programmes | Run a cohort through a term and its debrief | One term |
Interested in any of these? Write to hello@synthetic.school and say which.
Pilots are free or at cost while Synthetic School is validated, and every result is published.