Deploy
A fixed monthly cost, no per-message bills, no token surprises. Every installation is different, so the figure comes in a fixed proposal after a 20-minute call.
See a real agent working in Teams and tell us where the team loses time. We leave with one or two candidate roles.
Who: whoever manages the team
1 to 2 weeks: process map, roles ranked by hours freed, data and security assessment. At the end, a fixed proposal for the pilot.
Who: managers in 1 h interviews; us for the rest
Platform installed on your VM, Chief of Staff and first agent in Teams, KPI measured, governance handed over.
Who: IT: one VM and one M365 consent; managers: approve with a tap
Monthly, no lock-in. New agents hired by conversation, monthly review, usage and cost brief.
Who: your managers, by chat; us for operations
Three ways to start
All quoted on request. We pick one together on the call, depending on team size and what you already run.
on request
Audit + 4 weeks · fixed price
Most common
on request
Monthly, fixed cost, no lock-in
on request
For groups, several entities or specific IT requirements
Seats
A seat is a fixed monthly model subscription (Claude or OpenAI) that powers the agents. You don't pay per message or per token: the cost is the same in January and in July.
Catalogue
What we do, how long it takes and what we deliver. Figures in the proposal.
| Code | Offer | Duration | Deliverable |
|---|---|---|---|
| S1 | Readiness audit | 1–2 weeks | Process map, 5–10 candidate roles ranked by hours saved, data and security assessment, fixed pilot quote. |
| S2 | Pilot: first coworker | 4 weeks | Platform installed on your VM; Chief of Staff + one role agent live in Teams; KPI baseline; manager training (2h); governance pack included. |
| S3 | Team expansion | per agent, 1–2 weeks | New agent hired by chat by your Chief of Staff: charter, tools, schedules. |
| S4 | Managed operations | monthly | Updates, model upgrades, watchdogs, backups, healthchecks, monthly review, usage and cost brief. |
| S5 | Seats | monthly | Claude / OpenAI subscriptions that power the agents — fixed cost, no per-token billing. |
| S6 | Fractional AI operations lead | monthly | A named person who attends ops meetings, prioritises the agent backlog and reports ROI. |
| S7 | Training: managing AI coworkers | half-day | For managers: briefing agents, approving actions, conventions, memory, what not to delegate. |
| S8 | Governance pack | 1 week | Agent governance policy adopted by the client, approval matrix, compliance records (DPIA template, AI systems register, log retention ≥ 6 months) and the Portuguese worker-information notice (Lei 13/2023) ready before go-live. |
Terms
Programmes open and close — we confirm with you what is open. We do not guarantee approval. See compete2030.gov.pt and iapmei.pt for current status.
FAQ
Because every installation is different: number of agents, integrations, infrastructure, compliance requirements. In 20 minutes we understand the context and send a fixed proposal, excluding VAT, with no surprises afterwards.
The pilot is fixed-price and 4 weeks. After that, managed operations are monthly with no lock-in; annual prepayment gets a discount.
50% on signature, 50% when the first agent goes live. Monthly fees by direct debit or annual prepayment. Public-sector clients on their own procurement terms.
A fixed monthly model subscription (Claude or OpenAI) that powers N agents. A typical SMB runs 3 to 5 agents per seat. The Chief of Staff tells you when it is time to add one. Seat cost is stated in the proposal and in the monthly brief.
You keep everything: repository, containers, charters, data, governance policy. The hand-over is documented.
Anthropic (Claude) and OpenAI / Azure OpenAI in EU regions, under their DPAs. Open-weight models on your own server are on the roadmap for those who cannot send data out.
Programmes open and close monthly, so the site promises nothing. See the “How to fund it” box and talk to us: we confirm what is open at the time.
Pilot in 4 weeks
A 20-minute call, a real agent at work, no slides.