The founder operations platform for AI-native companies
Stop being your company’s human API.
Today, the founder is the memory layer. Crew One makes that memory operational.
No card. We onboard Founding 100 companies directly.
Your company’s operating context lives in one head. Yours.
Conversations in email. Plans in Notion. Decisions buried in Slack. Customers in the CRM. Every tool holds a piece. Only you hold the whole.
So work keeps returning to you — not because nobody else can do it, but because nobody else has enough context to move it forward safely.
At first, that looks like speed. Then every new customer, hire and tool sends another request for context back through you.
You explain the business to ChatGPT.
Again.
You brief Claude on the same customer.
Again.
You re-onboard every new tool with the same context.
Again.
You re-derive a decision somebody made last quarter.
Again.
AI makes output abundant. Company judgement stays scarce.
Anyone can generate a draft, analyse a spreadsheet or automate a task. What remains hard is knowing what matters, what has already been decided, which exceptions count and what should happen next. That is the operating judgement Crew One preserves and puts to work.
Crew One turns the judgement in your head into a system the company can reuse.
Start with the work that keeps pulling you back into the loop.
What changed overnight, what needs your judgement and what can wait — pulled from your inbox, calendar and connected tools.
Capture the decision, route the next actions, update the relevant systems and bring back anything that needs your approval.
Research the account, prepare the meeting, draft the follow-up and update the CRM. External sends stay held for your approval.
Start with one workflow. Build an operating system over time.
Most AI remembers what was said. Crew One records what was decided.
Every approval becomes precedent. Ask “what did we decide last time?” and get the answer, with the rationale, the crew, and the outcome.Context tells you what happened. Decisions tell you why it was allowed to happen.Six rows from the live trace: the actor, the action, the reason, the outcome, and a path back to the rule that produced it. This is the surface most products skip. It’s the one that lets a team trust autonomy.
Decision trace
· Last 24hNone active in 90 days. Downgrade queued for approval.
Only mutual slot before Friday. Inside the calendar autonomy rule.
GTM cited the Q3 Acme precedent. Legal cleared the revised clause.
Above the £30k threshold. Legal cleared the renewal terms at 08:02.
Used your approved framing: “Lead with the customer, not the demo.”
Previous first-touch proposals were rejected. Last-touch remains active.
You decide what it can do without you.
Set approval rules for money, external communications, policy and company data. Crew One drafts, routes and asks until you deliberately promote a repeated action into a trusted routine.
Every change to its autonomy is logged, inspectable and reversible.
Crew One earns autonomy. It never assumes it.
Approve Datadog renewal · £42k / year
Auto-renews Friday. Legal cleared terms; the 12% YoY increase matches the escalator you signed in 2024. Two cheaper alternatives surfaced; neither matches your coverage.
Per connection, per crew. Tighten the tier on the way in; widen it as the crew earns autonomy. Every promotion lands in the trace.
One Chief of Staff. Eight specialist crews. One operating memory.
Brief your Chief of Staff like you would brief a colleague. It routes the work to the crew that owns it, carries context across handoffs and returns only the decisions that need you.
Finance, GTM, Product, Marketing, Legal, RevOps, Operations and Executive Assistance all work from the same decision record.
Every approval, edit, rejection and exception can become precedent the whole company reuses.
Tools execute commands. Crews carry work forward.
Every crew draws from the same operating memory. Add crews as the company grows.
Every correction improves the next run.
When you edit, reject or explain an exception, Crew One captures the reason and proposes an update to the relevant playbook.
Once approved, the next similar task starts from that precedent instead of from scratch.
“The annual prepay calculation is wrong on Acme. It should be 12%, not 10%. Discounts are tiered by deal size.”
Mercury quote drafted at 12%.
Crew One cited the approved Acme correction and attached the source Decision Trace.
✓ Correct on the first run.
You teach the judgement once. The company can reuse it.
One operating thread, wherever the work happens.
Ask by voice, review in Slack, approve by email and continue on the web. Each surface returns to the same crew, Decision Trace and operating memory.
Chat answers. Automations follow rules. Agents act. Crew One coordinates.
It coordinates all three around your company’s judgement. Each alternative is useful; the missing layer is continuity — who remembers why the last decision was made, carries that precedent into the next task and shows you what will happen before it happens?
Remembering is useful. Putting the right precedent to work is the difference.
Least privilege by default. Every action on the record.
Give each crew access only to what it needs. Require approval where the stakes are high. See what acted, why it acted and which rule allowed it.
Your data remains yours. Export or delete it at any time. Private company data is not used to train shared Crew One models.
Your operating manual. Updated from approved decisions.
Crew One turns decisions, exceptions and repeated handoffs into a living ‘How Your Company Works’ playbook. Every change links back to its source Decision Trace. Search it, cite it, compare versions and roll back anything that no longer applies.
How your company works,
regenerated each Monday.
The longer you run it, the thicker your playbook.
The thicker your playbook, the smarter the crew.
Your business runs on you. That’s the bug.
Companies still run on org charts. Information runs up through middle managers. Decisions run back down. Most of the work is people moving context between apps.
A new shape of company is being built. Smaller, faster, recursive. The work happens in loops, and each loop teaches the next. Memory compounds. Headcount stays small because the company no longer runs on headcount.
Crew One is the first hire for that company.
We’re user zero.
Crew One has run our own company from day one: every decision logged, every crew handoff, every rough edge hit before it reaches you. The product on this page is the one we run the company on. It is not a mockup.
The Founding 100 is how we bring the first hundred companies on — directly, with the founding team.
Operating capacity without another full-time hire.
Priced against founder time and fractional support — not another chat subscription.
Every plan includes the full Crew One operating system: specialist crews, Chief of Staff, Decision Traces, approval controls and operating memory. Plans differ by monthly operating credits, background cadence, connected surfaces, collaboration, governance and support. Top-up credits are available for heavier months.
Questions that come up first.
The ones we hear most, answered straight.
Crew One is a Founder Operations Platform for AI-native founders and small teams. It captures the decisions, approvals, workflows and company context that today live in your head, turns them into private operating memory, and lets a crew of specialist agents move the work forward — with you approving what matters.
Solo founders, AI-native founders and small startup teams that need more operating leverage before hiring a full team. It suits companies where the founder is still the bottleneck for decisions, approvals and context.
An AI Chief of Staff coordinates work across strategy, operations, finance, GTM and product, and keeps track of what was decided and why. Crew One uses that model, then goes further — building operating memory and Decision Traces so the coordination compounds over time instead of resetting each session.
Chat assistants retain conversation context. Crew One creates an operating record: decisions, rationale, approvals and outcomes that specialist crews can reuse across future work — it does not just recall what was discussed. It coordinates what happens next.
You can assemble agents, automations and a vector database. The continuing work is permissions, evaluation, approval logic, decision provenance, cross-domain routing and maintenance — Crew One packages that operating layer so you can spend time building the company rather than maintaining its AI plumbing.
Good. Crew One prepares, coordinates and routes the work. You set the approval rules for money, external communications, policy and company data — as repeated actions become predictable, you can widen its autonomy deliberately. Nothing needs to become automatic simply because it can be.
Yes. Crew One connects the models and tools you already use through one operating memory, routing layer and approval system — it is the coordination layer, not another isolated assistant.
Each integration uses scoped permissions, with read-only access wherever possible. Approval rules are separate from technical access, and every action is logged with its rationale and source — your data remains yours. Export and deletion controls are built in, and private company data is not used to train shared Crew One models.
It starts with what you explicitly tell it and the sources you choose to connect — useful precedent accumulates as real decisions are approved, edited or rejected. The value compounds with use; Crew One does not pretend to know what you have not taught it.
Wake up to a briefing, not an inbox.
Run it with a Chief of Staff that knows what was decided, brings you what needs judgement and moves the next piece of work forward.
Built for founders who want to run on systems, not heroics.