A clinic directory patients can actually trust.

Rated Clinics is our own UK clinic directory. We built a five-dot score people can read at a glance, grew it from about 242 to 4,300+ listings with CQC register imports, and run it with daily AI agents.

Rated Clinics is a product we built and run ourselves. The job was simple to say and hard to do: give people in the UK one place to find a clinic and a clear, fair signal of whether it is worth trusting. It had to cover far more clinics than a hand-built list ever could, stay accurate as the registers change, and keep growing without a large team behind it.

Most clinic listings online are either paid placements or star ratings with no method behind them. A patient choosing where to have treatment has little to go on beyond a nice website and a handful of reviews. We wanted a directory that showed its working, and that meant three problems at once: getting enough clinics listed to be useful, scoring them in a way people understand, and keeping thousands of records current without doing it all by hand.

We started with the question a patient actually asks: can I trust this place? Reviews answer part of it. Safety, results and how openly a clinic explains its prices and practitioners answer the rest, and that part was hard to find in one place.

So before we designed a single screen, we worked out what a trustworthy signal would need to include and how it would be weighted. That became the Rated Clinics Score. Everything else in the product, from the listing page to the search results, is built around showing it clearly.

This is our strategy first stage in practice: decide what the product is for, then build the smallest thing that proves it.

A directory with a few hundred clinics is a list. A directory with thousands is somewhere people come back to. Before the import work, Rated Clinics had about 242 listings.

The Care Quality Commission publishes a register of the providers it regulates. We built imports that bring the CQC register into Rated Clinics as listings. That took the directory from about 242 listings to more than 4,300, roughly seventeen times the size.

Because the listings come from the register rather than from copy and paste, the directory stays tied to its source instead of relying on memory and spreadsheets. It is the same approach we now use on directories and marketplaces for other sectors: start from public data, then add the layer people cannot get anywhere else.

A score is only useful if people can read it in a second and check it when they want to. We chose five dots rather than a percentage or a star rating, so it reads as a signal rather than a review average.

The Rated Clinics Score is weighted across four areas:

• Safety, 30%. The things that protect a patient.
• Outcomes, 30%. Whether the treatment does what it should.
• Transparency, 20%. How openly a clinic explains itself.
• Experience, 20%. What it is like to be a patient there.

Safety and outcomes carry the most weight on purpose. A clinic can have a beautiful waiting room and glowing reviews, and still score lower than a plainer clinic that does the important things well. Being open about the weights keeps the score honest: anyone can see why a clinic sits where it does.

Thousands of listings create thousands of small jobs: clinics to contact, records to tidy, pages to write, posts to share. We did not want to hire a team to do them, and we did not want software publishing anything without a person seeing it first.

So Rated Clinics runs Claude agents every day, each with one job:

• Outreach. Drafting outreach messages.
• Data triage. Sorting new and changed records so the odd ones reach a person.
• SEO content. Drafting pages that help people find the right clinic in search.
• Social drafting. Preparing posts for the directory's channels.

Nothing goes out on its own. Every piece of work lands in an approval queue, where we read it, edit it if needed and approve or bin it. The agents do the first draft overnight; a person still signs off in the morning. We wrote more about the pattern in AI agents that work overnight, with a person still signing off, and it is the core of our AI agents and automation work.

The product is built in Laravel and Filament, with Inertia and React on the public side. Filament gives us a fast admin for running the directory; React keeps search and the listing pages quick for patients.

Since the CQC import, the directory has grown from about 242 to more than 4,300 listings, and the agents run daily. [RESULT: organic search traffic since the import] [RESULT: clinics claiming or updating their listings] [RESULT: patient enquiries passed to clinics]

Search is where most people will find a directory like this, so the next stage is SEO and AI search: making sure the right listing turns up when someone asks Google, or asks ChatGPT, for a clinic near them. If you run a clinic and want the same thinking on your own site, see our work for healthcare and aesthetics.

You can see Rated Clinics live at ratedclinics.co.uk.