What does my business actually look like at month 6?
Your AI employees handle the daily volume — outbound, support, marketing, invoicing, morning brief. You review and approve what matters, ignore the rest. You own the full setup and can run it yourself at $99/mo or stay on retainer. Most clients say month 3 is when it clicks: Monday morning stops being the worst part of the week.
What if I'm already drowning — will this add to my plate?
The first two weeks need about 10 hours from you: two voice sessions, a data walkthrough, and a tools audit. After that, we run the install. Your weekly check-in is 30 minutes. The whole point is to reduce your plate, not add to it.
Can I cancel during the six months?
The six months are binding. You can pause, not cancel mid-contract. We're direct about this because real voice training and playbook work takes time — shorter engagements produce weak AI employees that reflect badly on your brand. We'd rather turn you away than under-deliver.
Who is this actually right for?
Founder-led companies, $1–10M revenue, where the founder is still doing too much. If you're a solopreneur happy tinkering, self-serve at $99/mo is better value. If you have a 200-person org with a real ops team, Greenfield is the right call. The Agency tier exists for the middle — too big to do it yourself, too small to hire a full ops team.
Why only three clients per quarter?
Because we do the install ourselves — we don't hand you a template and walk away. Each engagement needs real voice capture, real playbook work, and a real human on our side. Three is what we can do properly. When we're full, we're full.
What AI model do my employees run on?
Model-agnostic by design. Default routing puts Kimi K2 (Moonshot) and Groq on the bulk of operational traffic; Claude, GPT, Gemini, or a local model plug in when you prefer them or compliance requires it. You bring your own API keys, or run on ours. The smart router falls back across providers so your install holds 99.5%+ AI uptime even if one model is down.
What happens to my data?
It lives in your own private install — your own database, your own brain. Nothing is shared with other customers. We never train on it or sell it. At month 6 you receive a full export of everything: playbooks, voice settings, knowledge graph, and audit logs. You can keep running it yourself at $99/mo with zero loss.
Can I see the actual prompts and reasoning?
Yes — every AI action is inspectable from your dashboard. Click any draft and see the full reasoning trace, the documents retrieved from the brain, the connectors called, and the response. Nothing is a black box.
Do you support EU data residency?
Yes — we can pin your install to a Scaleway Paris region. Default is EU-West. We do not use Cloudflare or third-party CDNs for your data.
How do you measure quality?
Voice fidelity score (target ≥0.82, measured weekly against your held-out writing samples), approval rate per AI employee (target ≥80% by month 3), and the outcome KPIs we set with you in week 1. Below 0.75 fidelity: we retrain. We report all of this to you weekly.
