I built Kai to catch the work businesses keep missing.
I'm Connor. KaiCalls answers the phone and keeps the lead moving. Kai CMO is the marketing system I'm building next. For work that does not fit inside a product, I map and install one repeatable job at a time.
The phone rings. Kai answers before it becomes voicemail.
KaiCalls is a phone secretary for service businesses. He answers when your team cannot, captures why the person called, qualifies the opportunity, and sends the business a usable next step.
See how KaiCalls works
- 00:00Incoming call answered01
- 00:18Reason for the call captured02
- 01:07Next step identified03
- 01:34Owner receives the handoff04
Kai CMO
The marketing system I use to turn research and requests into briefs, reviewed work, publishing, and reporting without rebuilding the week from scratch.
Follow the buildThe hard part isn't getting AI to answer. It's getting it to finish.
A useful system needs to know where the information came from, what it is allowed to change, when it has to stop and ask, and how we can check the result.
That is the work behind Kai: one clear job, limits that make sense, and a result a person can check. I write about the failures because that is usually where the useful rule came from.
More about meGive me the job that keeps needing you.
I map the job, decide what the software can do on its own, add the places where it must stop, and connect it to the tools you already use.
- 01Free
Score the workflow
Find out whether the job is repetitive, measurable, and safe enough to hand off.
You get: Fit, risk, and what is missing
- 02$750
Blueprint the job
Map the trigger, decisions, tools, limits, approvals, failure cases, and build scope.
You get: Workflow map and acceptance tests
- 03From $5,000
Install and test it
Build one tightly scoped workflow into the tools the business already uses.
You get: Working path, testing, and handoff
Some implementations use Hermes Agent, open-source software by Nous Research. I do not own Hermes. The paid work is the workflow design, integrations, limits, testing, and implementation around the job.
What I'm building, what did not work, and what I changed afterward.
From builds to lessonsCompleted builds become plain-English notes and reusable blueprints. Courses and workshops come after the same pattern has worked more than once, not before.
Start with a workflow scoreOne build note when there is something worth showing.
A real system, the part that failed, and what changed afterward. No daily AI news and no pretend lessons from work I have not done.

