This is for the independent compounding pharmacy owner who keeps hearing "AI" and has no idea what it actually does for a place like yours.
I'm Aaron. I build software for compounding pharmacies. Not slides about software — the actual thing, running in a real lab. Below are seven plays I've either built or would build for a pharmacy this week. Plain English. No theory. Each one is a real headache, what software does about it, and the smallest first step you could take Monday.
You already know pharmacy better than I ever will. This is the other half — what the machine can take off your plate so you can do the part only you can do.
The problem. Every GLP-1 script that comes in, somebody on your team reads it, then picks the right formula from memory or a binder.
The AI move. A tool reads the incoming script the second it lands — pulls the drug, the dose, the route — and maps it straight to your compounding formula. No one types it twice. If the script is missing something or looks off, it flags it and waits instead of guessing. It's not magic. It's one boring job done the same way every time, before a human ever touches it. That's exactly the kind of job software is good at and people hate.
How to start. Open your last week of GLP-1 scripts and count the steps between "script arrives" and "tech knows which formula." That count is the thing worth deleting. You don't need to build anything yet — just see how many hands and screens it touches.
The problem. You compound a little of everything, and you can't tell which lanes actually make you money versus which ones just keep everyone busy.
The AI move. Software pulls your own dispensing history and sorts it for you — what you compound most, what reorders, what's seasonal, what one prescriber drives. You stop guessing which niche to lean into because it's sitting right there in your own data. The point isn't a fancy dashboard. It's that the answer to "what should we be known for" already lives in your system, and a machine can surface it in an afternoon instead of you reconstructing it from memory.
How to start. Pull a simple report of your top compounds by volume for the last year. Even raw, it'll tell you something. That export is the seed — once you can see the pattern, software can watch it for you going forward.
The problem. Your formulation list lives in one place, your inventory in a spreadsheet, your logs on paper — and the same information gets re-typed into all of them.
The AI move. This is the big one. Most of the manual work in a compounding pharmacy isn't pharmacy work — it's somebody copying the same thing from one screen into the next because the tools don't share. Software sits in the middle and moves it for you: a formula gets pulled, the inventory count drops, the log entry writes itself. One action, not three. The owner stops reconciling by hand at nine at night because the numbers were already in sync when the work happened.
How to start. Pick one piece of information you type more than once a day. Just one. Trace every place it has to go by hand. That trail is the first thing worth connecting — and it's usually smaller than it feels.
The problem. A few words on your website or intake form can draw the wrong kind of attention, and you're too close to it to see them.
The AI move. A tool reads your public pages and forms and flags the language that reads as a problem — the "same as the brand" type phrasing, the claims that don't belong. It's a second set of eyes that never gets tired and reads everything you've ever published, not just the page you remembered to check. Quick note so I don't overstate it: the FDA action on compounded GLP-1s right now is a proposed rule in an open comment window — not a ban, not final. Worth watching closely, not panicking over. A language scan is good hygiene either way.
How to start. Copy your homepage and intake form text into any AI chat and ask it to flag wording that could read as a claim. Free, five minutes, and it'll surface things you've stopped seeing.
The problem. Every clawback, MAC dispute, or audit request means digging up the same backup and writing the same letter from scratch.
The AI move. Software keeps the paper trail together as the work happens, then drafts the response when you need it — pulling the right records and filling the template so you're editing instead of starting from zero. The appeal that used to eat an afternoon becomes a fifteen-minute review. You're still the one who signs it. The machine just stops making you rebuild the file every single time.
How to start. Save your last three appeal or audit responses in one folder. Those are your templates. The next one, have an AI tool draft from them — you'll feel the difference immediately and it costs you nothing to try.
The problem. Patients fall off their compounded meds, and nobody has time to chase every refill by phone.
The AI move. Software watches who's due and reaches out for you — a text at the right time, a nudge before they run out, a flag to a human only when it actually needs one. Your team stops being the reminder service and the patients who would've quietly disappeared get caught. Refills that were leaking out the back door come back, without anyone adding a task to their day. The system does the chasing; your people handle the conversations that matter.
How to start. Look at one compounded med people are supposed to stay on and ask who lapsed in the last ninety days. That list is your leak. A simple automated nudge to that group is the easiest win on this whole page.
The problem. New patients and new scripts pile up at the front of the line because onboarding is all manual — forms, data entry, follow-ups for the missing piece.
The AI move. Software takes the intake in clean, checks it for what's missing, and chases the gap itself before it ever reaches a person. By the time a human looks, the file is complete. No more half-finished onboardings sitting in limbo waiting on one missing field. The front of your pharmacy stops being where things go to wait. New work moves through ready instead of needing a person to babysit each step.
How to start. Find the one missing field that holds up the most intakes. Just that one. Automating the chase for that single gap will unclog more than you'd expect.
None of these are someday-tech. They're the kind of thing I build for compounding pharmacies now, one at a time, starting with whatever's costing you the most.
I send a play like this every week — a real one, the actual useful thing, not a teaser. If something here hit a nerve, reply and tell me which one. And if you ever want to see what one of these looks like running in a real pharmacy, come find what I build.
— Aaron
One AI play for your compounding pharmacy. Every week. · relayosmedia.com