Workshop series · 6 sessions
The Agent-Native Agency workshop
Build an agency that runs on agents. Whatever you run today — Claude Code and Codex agents, n8n flows, a few scripts — we bring it onto a platform you own, then build what’s missing. Six live sessions, step by step. Build along and leave with your own.
One registration covers all 6 sessions.
Hosted by Eugene Vyborov
Founder of Ability AI · creator of Trinity
Come ready to build along — 2-minute prep (optional): skim the Trinity docs and have the GitHub repo open.
The series
Six sessions, one agency.
Sequential and build-along: each session stands on the last, from what you run today to an agency that runs unattended on a platform you own.
What Trinity is, and your own instance running by the end
RecordingA tour of the production instance running our company on about 40 agents, then your own Trinity instance installed on DigitalOcean in about five minutes.
Bring your existing agents, then build the departments
RecordingYour existing agents and flows sorted into keep, wrap, or rebuild, one ported live onto Trinity, then the agency laid out as departments of agents, each owning its own metrics.
Give your agency one shared brain
RecordingThe shared knowledge base every agent reads, plus the data layer that collects what agents produce and an interpretation layer that turns numbers into business conclusions. This is what keeps every department consistent and on-brand.
Orchestration: one agent to run them all
RecordingInstead of juggling a dozen agents, drive the whole agency through one orchestrator — you talk to one agent in one place, and it runs the rest. The org design and hand-offs that make a single point of control actually work.
Workspace: the interface for your team
Recording soonSet up Workspace — the non-technical interface where your team (and, when ready, your clients) works with the agents: chat, files, voice, canvas, work loops, chat history, each person scoped to the agents they need.
Operations & the commercial model
Next sessionRun the agency unattended and actually see it running — what’s healthy, what broke, why — with self-recovery and clear signals for when to step in. Then the question every agency founder has first: how you sell and bill agent-run work — one sovereign, dedicated instance per client, what to charge, and margin versus headcount.
Built on Trinity
Every session runs on Trinity — the open-source runtime for production agents, created by Ability AI and used to run its own company. It’s Apache 2.0 and free to self-host on infrastructure you control, so the agency you build stays yours: your agents, your data, your servers.
Who it’s for
For the people who run the agency.
Bring a laptop and the agency you have — or the one you’re starting.
Your host
Who’s building.

Eugene’s lab works out how companies should run on agents, then puts them there — and runs his own on them. He builds live every week and publishes what works.
Where this starts
The big picture, in one hour.
This series is the hands-on build of the webinar below — everything you run today brought onto one instance you own, run for clients without becoming their IT department. Watch it for the model; join the series to build it.
Save your seat
Join the series.
One registration, all 6 sessions. Bring your laptop and build along — and if you miss one, the recording lands in your inbox.



