A hundred hands for the work behind your work.
We research your market, build the systems and AI agents your team runs on, and test that it all holds up. Then we hand it over so your team keeps the leverage.
hekaton · Greek for one hundred. The Hekatoncheires were the hundred-handed giants who did the heavy lifting that won the war of the Titans.
Start with what's breaking.
Six problems we see every week, across five practices: research, operations, automation, quality and growth. Pick the one that sounds like yours. Each has a fixed scope, a deliverable you can check, and a price you know before we start.
Hundred-Hand Audit
Two weeks mapping how work actually moves through your business. Every recurring task gets scored for automation, delegation or testing, and you get a ranked build plan with costs.
- Workflow map of your team
- Task-by-task leverage scores
- Ranked build plan with costs
- Readout with your founders
Market Read
Before you build, hire or launch, find out who is buying and why. We run customer interviews, tear down your competitors and size the market, then turn it into positioning you can test.
- 10–20 customer interviews
- Competitor teardown
- Market sizing
- Positioning tests
Operator-in-Residence
A senior operator embedded with your leadership. We set the operating rhythm, close loose ends, and turn what lives in founders' heads into systems the team runs on its own.
- Weekly operating cadence
- SOPs and playbooks
- Hiring and onboarding
- Board and investor prep
Agent Builds
AI agents and automations for the repetitive middle of your business. They propose, your people approve. Every build ships with documentation and training, so your team runs it without us.
- Inbox and intake triage
- Automated reporting
- Knowledge assistants
- Team training and handover
QA & Evals
Testing for your website, product and AI systems. We write the test plans, run regression before every release, and build evals that check your AI gives the right answer, not just an answer.
- Test plans and cases
- Regression and UAT runs
- AI agent evals
- Accessibility checks
AI Search Visibility
People ask ChatGPT, Claude and Perplexity before they ask Google. We make sure you show up and get cited, then build the content engine that keeps you there.
- GEO and SEO audit
- Citation strategy
- Content system
- Visibility dashboard
How many hands are you missing?
A rough sense of the time your team spends on repeatable work, and how much of it could come back. The audit replaces this estimate with real numbers.
Estimate only. Assumes a 40-hour week and 46 working weeks a year.
Plan first. Ship in pieces.
Every phase ends with something that works on its own. You can stop after any of them and keep what you've paid for.
Audit
Map how work actually moves. Find the drag. Size the opportunity.
Blueprint
A written build brief: scope, stack, costs and success criteria you sign off on.
Build
Small, testable milestones. Each one runs before the next one starts.
Run & hand over
We operate it with you, document it, and train your team to own it.
Five rules we don't bend.
Propose, then act.
Systems suggest, people confirm. Automation earns its autonomy.
Ship before scaling.
Useful in weeks. We grow what works and drop what doesn't.
Own your stack.
Open tools, your infrastructure, full documentation. No lock-in.
Charge for outcomes.
Fixed scopes tied to results. Hours are our problem.
Teach as we build.
Your team should need us less every quarter.
Your build, live, any time.
Every client gets a dashboard showing what's shipped, what's in progress, what's waiting on you, and the hours each system has returned.
Sample data for illustration.
Before you ask.
Who do you work with?
Founders and operators at ventures of roughly 5 to 100 people who have outgrown doing everything by hand but aren't ready for a full ops team.
Do we need to be technical?
No. We write the plan in plain language, build the tools, and train your team to run them. You review outcomes, not code.
What happens to our data?
It stays in tools you own. Where it makes sense we self-host, and we never train models on your data.
Can we stop partway through?
Yes. Every phase ships something that works on its own. Stop after any milestone and you keep everything built so far, with documentation.
Tell us where the work piles up.
We reply within one business day with a call slot or an honest "not a fit".