Custom & Self-Hosted AI Systems

Workflow: AI Content Machine

Before: Regular blog posts are one of the best ways for a local clinic to win new patients — but writing and publishing one a week is half a day's work no solo owner can keep up.

Outcome: The clinic keeps publishing without a writer, an agency, or more than a few minutes a week. Its presence on Google keeps building on its own.

System: Every Monday it finds the search topic most likely to bring in customers and AI writes the full post. The owner reads it and clicks approve, and it goes live on its own. The owner's whole job is one click.

Workflow: Weekly Search Health Score

Before: A clinic owner has no single view of how findable they are. Google rankings, map results, site health, and whether they show up in AI answers like ChatGPT all sit in separate tools — or nowhere.

Outcome: The owner gets one weekly read on how findable the clinic is — no tool-hopping, no guessing. Problems surface on their own, progress is easy to watch, and the owner can see whether Google is actually bringing in patients.

System: Every week it checks four things — how the clinic ranks on Google, where it shows up on the map across the area, whether the site is healthy and fast, and whether it appears in AI answers like ChatGPT and Gemini — and rolls them into one 0–100 score with a clear to-do list. The owner sees one number and knows what to fix.

Workflow: Run the Clinic by Text Message

Before: The clinic's content, SEO, and social tasks involve many steps across many tools. Without something tying them together, each one means remembering what to run and doing it by hand.

Outcome: The owner runs the whole clinic from their phone. The multi-step work happens reliably in the background, and the owner just sends a text.

System: All the automated steps — writing content, checking search, scheduling posts — are wired together to run reliably on their own. On top of that, a chat assistant lets the owner run everything from a text: approve a post, start a search check, check on the clinic, or kick off new content, with no dashboards to open.

Workflow: AI Study Helper You Can Actually Trust

Before: Students and scholars need fast, reliable answers about Quranic Arabic, tafsir, and grammar. General AI tools make things up and can't point to real sources, so they can't be trusted for serious work.

Outcome: Get scholar-grade answers in seconds instead of hours of looking things up by hand. Because every answer shows its source, you can check it and trust it.

System: Ask about a word, its grammar, its tafsir, or where a topic shows up elsewhere, and get an answer drawn from a large library of trusted Islamic scholarship. Every answer comes with the exact source it's based on — and it won't answer without one. So unlike general AI, it doesn't make things up.

Workflow: DVLA Fitness-to-Drive AI Assistant

Before: Drivers with medical conditions often don't know whether they must tell the DVLA or can sign off themselves. The answer depends on the exact condition — and general AI tools get it wrong far too often to trust.

Outcome: A documented, verified benchmark: 88% correct on DVLA clinical questions, against the best general AI's 46%. Built and benchmarked, ready for whenever the client chooses to switch it on.

System: An assistant that answers strictly from the DVLA's own fitness-to-drive guidance, then tested head-to-head against the leading general AI tools on a set of real clinical questions. In that benchmark it scored 88% correct against their best of 46%. It's built and benchmarked — going live is a decision for the client.

Workflow: Private AI, Running In-House

Before: Renting AI from a big cloud provider would mean unpredictable bills, handing sensitive religious text to a third party, and speed limits that make overnight runs impractical.

Outcome: New commentary keeps reaching tafsirtool.com, fast enough to cover the whole library, at the cost of a graphics card instead of a never-ending rental bill.

System: The AI runs entirely on one machine in-house, working on eight passages at once. It splits the text cleanly so no verse is ever cut in half, translates it, and tidies the result — all without sending anything outside. 15 classical scholars are already covered in the live product.

Workflow: Every Translation, Checked Automatically

Before: Translating fast is useless if mistakes slip through. People could never review every line by hand fast enough, and bad output would quietly pile up.

Outcome: Every line is checked, so only translations that pass reach scholars and readers on tafsirtool.com.

System: After each translation, a separate AI reviews it against the original Arabic, records a clear pass or fail, and sends anything weak back to be redone. Nothing reaches readers until it passes.

Workflow: Automatic Red/Amber/Green Safety Alerts

Before: Deciding whether to call the surgeon now or wait can't hang on gut feel — that's inconsistent and slow. A safety call this important needs clear rules that always fire the same way and can't be talked out of.

Outcome: Safety alerts that fire the same way every time, leave a clear trail, and can be reviewed by a Clinical Safety Officer without reading any code.

System: Reads the bedside observations and gives a Red, Amber, or Green result using fixed safety rules — no guessing, no AI making it up. The same readings always give the same answer, and every alert scenario has a test proving it still works. The rules run on the server, so the result can't be faked or skipped from the app.