Aurora CapacityOS runs supervised AI capacity inside your clients' systems. Agents do the operational work on one of four desks — Campaign, Retention, Insurance, Financial — the right human approves every step that matters, and a capacity ledger proves the hours and outcomes delivered. The pilot is how a firm finds out what that's worth on its own work: we pick one workflow, put it on three of your real clients, and run it together for 30 days.
A pilot only works if both sides invest. You get the desk running on real client work, set up personally by the team that built it, with a thirty-minute review every week. We get the kind of feedback you only get from a firm using the thing every day. At the end of the 30 days the ledger shows exactly what was delivered — work items completed, decisions captured, outcomes — and that ledger sets the number for a monthly capacity retainer if you want to keep it running. Or you walk away clean. No contract either way.
Five screens from Aurora CapacityOS as it runs today — the Marketing Desk on sample data, honestly labeled, plus the page your client sees on their phone.
Your morning view. Capacity delivered this month sits next to what's running now and what needs a hand — and two actions are paused at the approval gate. On the right, a decision packet spells out exactly what happens if you approve, and what can only partly be undone. The numbers are sample data, honestly labeled.
Every workflow states its terms up front: when it starts, which systems it works inside, and where you stay in control — nothing is sent on its own. The capacity it gives back, about 2.5 hours per client per month here, carries its assumption in writing. Even the status pills are honest: demo on sample data, or available in pilot.
Behind each workflow is a team of specialist agents — Coordinator, Researcher, Strategist, Writer, Reviewer — each with a "may" list and a "may never" list. Below them, the red lines: enforced in code, so an agent that tries one is blocked, not just discouraged. The first line reads "Never publish or send without the client's approval."
Each client gets their own cockpit: their record, their connected systems, and exactly how Aurora is allowed into each one. Here that's Google Ads and Google Analytics — both wearing their SAMPLE chips — with delegated access set up per client and a revoke button beside every connection.
This is what your client sees — on their phone, no login, no new tool. Two budget changes with the reasoning in plain language, the evidence one tap away, and two buttons: approve or request changes. Even the savings figure says what it is — a demo figure from sample data, not a real account.
Thirty days of real work, not thirty days of trying it out. We bring the desk and our time. You bring three real clients and honest reactions. Anything less and we both wasted a month.
One workflow, run properly, by the team that built it.
Without this, neither of us gets value out of the thirty days.
Aurora is an AI solutions studio working out of Stuttgart. The team builds one platform — Aurora CapacityOS — supervised AI capacity inside your clients' systems, organised into four desks: the Campaign Desk for paid campaigns reviewed and proven, the Retention Desk for revenue at risk and save plays, the Policy Desk for same-day certificates and prepared renewals, and the Advisory Desk for client onboarding and identity checks. On every desk the same rule holds: agents do the work, a human approves what matters, a ledger proves what was delivered.
We build for firms where the work is the business — marketing agencies, insurance brokers, and financial-services firms. The difference between an owner who sleeps well and one who doesn't is the kind of problem we care about. The studio is EU-hosted end to end, because the firms we want as customers don't want their client data scattered across providers they can't name.
That's also why the pilot cohort stays small. You're the firms we'd want as long-term customers, and the firms whose feedback we trust enough to ship from. We run Aurora's own desks on Aurora's own work — what we'd recommend, we already run in production.
Easiest: thirty minutes whenever it suits — on video, or in person if you're near Stuttgart. We walk you through the desk that fits your firm in real depth, you ask whatever you want, and we decide together whether the pilot makes sense for both sides.
Aurora