In 2018 a small group of us at the University of Exeter ran a project to map the actual grade distributions of optional modules in Computer Science. We had access to an unusually candid set of staff who shared aggregate data, and the result was a little dashboard that students used to make better module choices.

The pilot worked. Students got better grades. We didn't monetise it. Life happened.

Eight years on, the same problem still exists at every UK university — only more so. Module choice is high-stakes, the available signal is poor, and the data that would make it tractable is scattered across institutional FOI archives. Building it properly is now our full-time job.

What we're building

A paid AI advisor for UK students, trained against a proprietary, normalised dataset of module-level outcomes. It answers the questions students actually ask: which optional modules give me the best shot at a first? If I'm aiming for a 2:1, which combination minimises risk? Which modules have shifted lecturer recently?

You can read more about how we build the dataset or what we publish and what we don't.

Why this is different to a league table

League tables rank universities. We rank decisions within a university. That difference matters: most students aren't deciding between three universities by the time they're picking final-year modules — they're deciding between five modules inside one. That's the decision point we're built for.

We're starting with module choice for existing students. University and course comparison for prospective applicants is on the roadmap, but the existing market there is crowded; the module wedge is uncontested.