work / case studies / pharmassist

PharmAssist: mapping the AU pharmacy landscape

PharmAssist is a prospecting map of Australian pharmacies, built on Supabase and OpenStreetMap data. It's for pharmacies and pharmacy-adjacent businesses trying to understand the AU pharmacy landscape — where pharmacies sit, at a level of coverage a spreadsheet or a manual Google Maps search can't easily give you, within the limits of what OpenStreetMap's crowdsourced data covers. It's one piece of the broader pharmacy product I'm building under the PharmAssist name — the map is the prospecting layer; honest, evidence-first supplement guidance at the counter is the other. As of August 2026 it's pre-revenue, in a founding-customer offer stage, and I'm being upfront about that rather than dressing it up.

The problem

There's no clean, current, queryable picture of the pharmacy landscape in Australia — where pharmacies sit, how they cluster, what's nearby them. If you want that picture today, you're stitching it together by hand: council directories, individual pharmacy websites, and a lot of manual map-scrolling. That's slow, and it goes stale the moment a pharmacy closes or a new one opens.

What was built

PharmAssist is a map interface over a Supabase database, seeded from OpenStreetMap pharmacy data for Australia. The point of building it this way — real geographic data, a real database, a map you can actually explore — rather than a static list, was to make the pharmacy landscape something you can query and see, not just read about. It's the healthcare-domain proof point in what I build: the same "ship it, use real infrastructure, don't fake the data" approach I bring to client work, applied to my own pharmacy background.

Where the human gates sit, and what was hard

OpenStreetMap is the best open source of pharmacy location data available, and it's also inconsistent by nature — it's crowd-maintained, so coverage and accuracy vary place to place. Treating that data as ready-to-use without a verification pass would have meant shipping a map that quietly lied in places. So the raw OSM import is a starting point, not the finished product, and I don't represent the map as more complete or more current than it is.

The same caution applies to what the map is for. This touches real pharmacy businesses, and any outreach that comes out of it — a founding-customer conversation, an introduction — goes through me first. I'm not running automated messaging off this data. That's the same rule I hold everywhere else in healthcare-adjacent work: a human reviews anything that touches a real business or a real person before it goes out, and I'd rather move slower than automate the part that can quietly go wrong.

Current status

As of August 2026, PharmAssist is pre-revenue and in a founding-customer offer stage — I'm talking to pharmacies directly rather than reporting numbers I don't have yet. I'm not going to publish user counts, revenue, or growth figures here because there aren't any worth publishing. If that changes, I'll update this page and say so plainly, the same way I'd want a vendor to tell me.

This is one entry in a longer list of things I've actually shipped, half-shipped, or am still deciding whether to keep — see everything I've built for the rest.

Work with me

If you're in pharmacy or healthcare and want AI work built with this same caution — de-identified data only, a human sign-off on anything that touches a real patient or business, no invented metrics — that's exactly the lane I work in.

Email meSee how I work with clients →

the build log

Get the build log.

One email when I ship something — a new AI system, a pharmacy workflow, a number from The 2040 Project.

No spam, no drip sequence, unsubscribe in one click.

The weekly dose: graded, sourced, 5 min.Subscribe