Wednesday, May 20, 2026 · written 15:00 · 14h 26m tracked · blog, handyman, agentic-system

Day 7: two meetings, a partner locked, the blog grows up

A high-leverage day. A paid client walkthrough of my agentic system, then a working session that locked the Handyman partnership and the plan. In parallel the blog got a real upgrade: discoverability, a downloadable starter kit, a public Building section, a newsletter, and multi-language in progress. Reconciled at day's end with the numbers and the evening in.

Written mid-afternoon, so this is the preliminary cut. The Rize numbers and the evening get reconciled later, the way Day 6 did.

Two meetings carried the day.

The walkthrough

The first was a paid call where I walked a client through my agentic system end to end. The same hour did double duty. The setup is now a downloadable starter kit and a second post on the blog, both written while the explanation was still warm. The most valuable part came at the end, off the build topic entirely: a clear, specific breakdown of how to use LinkedIn to reach the right people, in the right order, founder first. I wrote it down before it could fade.

The partner session

The second was the one I had been getting ready for. A working session with my local partner on Handyman, and it had a hugely positive return on the time. We left it with the thing locked: who does what, how we grow, and the smallest first slice worth building. He drives the first local signups, I build the minimal app. We walked the demo, talked monetization without flinching at the hard parts, and named the disintermediation risk out loud instead of pretending it away. I came away more sure of the partnership and the person than I went in.

The blog, from the family

The push on the blog today came from the bluntest source possible. My grandmother could not find the site and could not read it. She searched the name and landed on someone else, and there was no German or Portuguese to fall back on. So today the blog got the things that feedback demanded: structured data and a proper identity so search stops confusing me with strangers, and multi-language, in progress as I write. Plus the parts I wanted anyway: a starter-kit download, a public Building section with a per-project changelog, and a real newsletter instead of a placeholder.

Still to come

A third Jiu Jitsu lesson tonight. A little movement after a day at the desk usually pays. The full version of the day, numbers included, gets added once it is done.

2026-05-21 evening reconcile

The numbers are in. A long day. Code led at over four hours, with meetings and in-person time close behind, over five hours combined. Marketing still got crowded out by the talking, and on this day that was the right trade. The conversations paid more than another block of build would have. There was a third conversation I had not counted when I wrote this at three. A long evening call with an advisor in the family ran two hours and was the highest-return stretch of the day. It sharpened the Handyman thesis in a way I had been circling but not saying out loud: the marketplace is the visible product, but the engine underneath is reselling micro-financing, and the people who actually pay are not the homeowners or the workers, they are the lenders who want a channel to sell through. The property-management idea is a sibling to it, a second front door to the same marketplace, not a feature buried inside it. Homework fell out of that too: actually size the market and the incumbents before I tell an investor anything. A marketing contact replied. The hope is to hand him OneProposal's marketing so it does not stall while Handyman takes the time it is about to demand. The honest theme of the day: work is piling up faster than I can clear it. Three high-value conversations is a good problem, and it is still a problem. Then Jiu Jitsu, a little late because the call ran over. I rolled with the same partner from my first session. He has my size and far more time on the mat, so he controlled most of it, but I held my own and tried things, and he had already seen every one of them. Rolled with the instructor too, who was not breathing hard while I gave everything I had. The value is not the moves. It is the derivative: confidence in your own body, calm under pressure, the standing it gives you to back yourself and other people. Same pull as the bouldering gym. The community is the reason you come back.

2026-05-21 ~02:00, calling it

A late push after an already long day. I shipped the Handyman waitlist sign-up page and tried to wire it to the click-through demo. The page is up, but the link to the demo is not working yet, so that piece is unfinished. I finally killed the OneProposal engine bug the Railway crash had been causing, which had hung over the last couple of days. And in parallel I started building a private admin page for the blog, an internal documentation hub behind a real login. It is half-built: it deploys and the gate is scoped, but the DNS and the access setup are not finished, so the internal docs are not live yet. The honest read at 2am: I am running in loops, patience low, the work turning counterproductive. That is the signal to stop. The targets that mattered today were already done by the evening. The half-finished pieces are tomorrow's, with a clear head. Logging off.

Targets

Walk a client through my agentic system on a paid call
Working session with my local partner to lock the Handyman plan
Clear admin in the morning
Level the blog up from family feedback: discoverability and languages
Third Jiu Jitsu lesson in the evening

Wins

The blog grew up today. Shipped the discoverability work so search can tell me apart from the other people with my name, a downloadable starter kit for the agentic setup, a public Building section with a changelog, and a real newsletter. Multi-language is in progress as I write this.
Locked the Handyman partnership in a working session with my local partner. We agreed who does what and how we move: I build the minimal app, he drives the first local signups. We walked the demo, mapped monetization and the disintermediation risk honestly, and narrowed the first slice to a waitlist plus a register-and-browse app.
A paid client walkthrough of my agentic system paid twice. The same hour became reusable: a downloadable kit and a second post.
Got concrete, valuable feedback on outreach at the end of the client call. How to actually use LinkedIn to reach the right people, in order. Captured it before it could evaporate.
Cleared admin in the morning instead of letting it bleed into the build.

Losses

Family could not find or read the blog. My grandmother searched the name in a browser and landed on a different person with the same name, and the Portuguese and German versions did not exist. It took family friction to surface what the analytics already implied: no external reach, and one language only.
Marketing still is not running as its own block. The outreach feedback makes the gap obvious. It needs to start now, in parallel, not after the building is done.

Lessons

The most useful testers are the ones who are not technical. One message from my grandmother trying to open the site found two real gaps I had filed under later.
Work I do for a paid client is also content and a product. The walkthrough was the client's hour and the seed of a post and a kit. Build the thing once, let it pay more than once.
Selling the person matters as much as selling the product. The outreach advice was a reminder that the founder is the first thing an investor buys.

How did this land?

Comments

Loading comments…