How We Generated Custom Summaries for Every UK School in Just 24 Hours
When I started building ukschools.info, my vision was simple: make it easy for parents, governors, and teachers to find clear, concise information about every school in the UK. The Department for Education (DfE) already provides a huge amount of public data, but it’s often overwhelming, fragmented, and not written in a way that’s helpful for everyday users.
So, I asked myself: what if every school had a custom-written summary — easy to read, consistent in style, and generated automatically?
That idea quickly turned into an experiment with AI.
From Raw Data to Readable Insights
The DfE provides a school’s URN (unique reference number) along with useful metadata — including the school’s website. Using this, I wrote a PHP script that scanned the homepage and key pages for each school. The script collected the visible text and paired it with the DfE’s structured data.
This information was then sent to OpenAI’s API, which was tasked with producing a short, three-paragraph narrative about the school. Instead of just raw facts, the output became a human-readable summary: the type of school, its ethos, its community role, and highlights that matter to parents and stakeholders.
Scaling Up: 30,000+ Schools in a Day
Of course, writing one summary is easy. Writing 30,000 summaries is another story. Doing this manually would take years.
That’s where automation — and a little engineering creativity — came in. I built a script to run in 40 threads in parallel, each handling a batch of schools. The script automatically:
- Pulled the school’s homepage.
- Extracted the main text.
- Combined it with DfE data.
- Sent it to the AI model.
- Published the summary back into the database.
The result? In just 24 hours, we generated unique content for every school in the UK.
The total cost for API usage? Around £20.
Why This Matters
For parents, governors, and teachers, this means every school now has a clear, accessible summary instead of just a wall of raw statistics. For schools, it provides a level playing field: each is introduced with consistent, unbiased information, created from its own data and web presence.
For me, it’s also a demonstration of how AI can transform public data into something meaningful at scale. What would take teams of writers months (if not years) to produce was achieved in a single day with the right mix of scripting, data, and AI.
This is just the beginning. The same approach can be applied to other public datasets — GP practices, charities, local authorities — anywhere there’s a need to turn complex data into accessible knowledge.
Final Thought
AI didn’t replace human creativity here — it unlocked it. By automating the heavy lifting, we’ve created a foundation that schools, parents, and communities can build on. And all it took was one idea, a bit of code, and £20.