When your doctor tells you to monitor your blood pressure daily, you nod, buy a cuff, use it twice, and then it collects dust on the nightstand. I know because that's exactly what would have happened to me.
Last week I walked into my doctor's clinic in Berlin — Dr. Israel Tamayo, a Cuban internist in Prenzlauer Berg who I've been seeing for years — for a routine checkup. Nothing dramatic. Blood pressure came back a bit high. Not alarming, but enough for him to say: "I want you measuring this every day, morning and evening."
And that's when I did what most people wouldn't: I delegated the entire workflow to my AI assistant.
"Asere, you have more work now"
Asere is my AI-powered digital assistant. He runs 24/7 on my infrastructure, handles my calendar, monitors my servers, manages deployments, writes code, and — as of last week — tracks my blood pressure.
Here's what I told him, roughly: "The doctor wants me to monitor my blood pressure daily. Help me buy the device, explain how to use it, create a tracking spreadsheet, remind me twice a day, and when I send you a photo of the reading, tell me what it means and log it."
That was the entire specification. No app downloaded. No account created. No subscription.
What happened next
Within minutes, Asere had:
- Researched blood pressure monitors — compared options, recommended one with a good cuff size for my arm. I picked one up at DM (the German drugstore chain) the next day. Done.
- Explained proper measurement technique — sit for 5 minutes first, feet flat on the floor, arm at heart level, no caffeine 30 minutes before, measure on bare skin. He didn't just dump a manual; he gave me the protocol in plain language.
- Created a structured log — a simple file with columns for date, time, systolic, diastolic, pulse, and notes. No over-engineered spreadsheet with 47 tabs. Just what matters.
- Set up daily reminders — morning and evening, timed to my schedule. Not a generic 8 AM / 8 PM. Timed around my actual routine in CET.
- Established validation rules — he won't log a reading taken right after coffee or while standing. When I tried to register a measurement I took after an espresso, he flagged it: "That reading is unreliable — caffeine elevates both systolic and diastolic temporarily. Take it again later at rest." And he was right.
The daily routine now
Every morning and evening, Asere pings me on Telegram:
🩺 Recordatorio: tómate la presión arterial. Dime los valores (sistólica/diastólica/pulso) y los registro.
I take the measurement, snap a photo or type the numbers, and send them. He logs the entry, tells me if anything looks off, and moves on. No friction. No app to open. No data entry.
After a week, the log has a dozen entries — structured, timestamped, validated. Morning readings, evening readings, notes about conditions. A clean dataset ready for my next doctor visit.
Instead of walking in and saying "uh, I think it was around 130-something a few times," I'll hand Dr. Tamayo a proper table with trends he can actually interpret.
Why this matters beyond blood pressure
This isn't really an article about blood pressure. It's about what happens when your AI assistant actually knows your life.
Asere didn't need a health app integration. He didn't need a Bluetooth API. He needed:
- Context — he knows who my doctor is, what my schedule looks like, and that I drink coffee
- Memory — he remembers previous readings and can spot trends
- Judgment — he rejected a bad reading instead of blindly logging it
- Initiative — the reminders aren't a calendar event I set; they're part of his operational routine
This is the difference between an AI tool and an AI assistant. A tool does what you click. An assistant anticipates what you need.
The uncomfortable truth about health apps
There are hundreds of blood pressure tracking apps. Some are genuinely good. But here's what they all require:
- Download the app
- Create an account
- Learn the interface
- Remember to open it
- Manually enter data (or buy their specific Bluetooth device)
- Hope they don't sell your health data
My workflow is: message my assistant on Telegram. That's it. The same interface I use for everything else in my life. Zero context switching. Zero new accounts. Zero data leaving my infrastructure.
What's next
I'm working on having Asere read the cuff display directly from photos — so I won't even need to type the numbers. He'll extract them from the image, validate them, and log the entry. We're also setting up a weekly summary that automatically compiles trends and flags anything my doctor should see.
The endgame isn't replacing medical professionals. Dr. Tamayo makes the calls. But between appointments, having an assistant that handles the boring operational work of "measuring, recording, remembering" means I actually do it. Every day. Consistently.
And consistency is the only thing that makes health monitoring useful.
Building Asere taught me that the best AI assistant isn't the smartest one — it's the one that removes enough friction that you actually use it. If you're curious about how AI-powered personal assistants work beyond chatbots, check out my other posts or reach out on LinkedIn.