You pay for governed agents that earn their autonomy — not for seats. An advisory agent costs little; value grows as agents climb the trust ladder and safely carry more of the work. You only ever grant the autonomy you've verified, so spend tracks the trust you've actually earned.
Aegis pairs every role on your org chart with an agent. So pricing follows the agents you put to work and the autonomy you grant them — not how many people log in. A pseudo agent that only drafts suggestions is inexpensive. As an agent demonstrates good work and earns higher postures, it takes on more of the routine load — and that's where the value, and the price, scales.
Value follows the trust ladder
pseudo · advisory onlysupervised · acts on approvalcontinuous_insight · human by exceptionEarned, never assumed. An agent only reaches a higher posture — and the value that comes with it — after it demonstrates good work under human authority.
Every plan is governed by construction and deployed in your environment. Pricing is tailored to your deployment — these tiers describe scope and capability, not list prices.
Prove it on one function.
supervised / shared_planningGovern a whole function.
continuous_insightYour whole org chart, governed.
delegatedHow pricing works: the Pilot is a fixed-fee, time-boxed engagement — a bounded first step you can approve without a long procurement cycle. Department and Enterprise scale with the number of governed agents and the depth of autonomy you grant, quoted to your deployment. We don't publish list prices because what you run — and what you let it do — varies by organization.
Here's the shape of the value, not a promise of results. As one agent climbs the trust ladder, the share of routine, low-risk work it can safely carry rises — so the effective cost per task falls. The percentages below are illustrative examples to show the direction of travel; they are not measured customer outcomes.
pseudo — drafts onlyThe agent suggests and drafts in chat; a person does every action. Useful, but it carries little of the load on its own.
supervised — acts on approvalThe agent does real work and a human approves each action. More of the routine clears itself, with a person still in the gate.
continuous_insight — acts within envelope, human by exceptionThe agent acts inside its operating envelope and a person steps in only for the exceptions. In this illustration an agent might safely handle a majority of routine, low-risk tasks — so the cost per task is at its lowest.
The logic, not a guarantee: higher earned autonomy means an agent safely carries more of the routine work, which lowers the effective cost per task. Where any particular function lands on this curve depends entirely on your work, your risk tolerance, and the autonomy you choose to grant. Any percentage or share shown here is an illustrative example — never a measured customer result.
Tell us nothing — tell yourself. The fields below open with example values — change each one to match your own organization and every total updates as you type. We just do the arithmetic in front of you. Nothing here is an Aegis price, quote, or promise of savings; it's your numbers, multiplied.
The annual labour value of the routine time you'd be putting under agent support — at the share you chose.
This is arithmetic on your inputs — not a result. It shows the size of the opportunity you're weighing, not a saving Aegis delivers or a return it guarantees. To turn it into a business case, weigh this figure against Aegis's fee and your own rollout effort — the fixed-fee pilot is designed to bound that cost, so you can test the return before committing. What an agent can actually carry depends on your work and the autonomy you grant.
Whether you're piloting one function or governing the whole organization, the foundations are the same. Governance, audit and open standards aren't upsells — they're how Aegis is built.
Operating envelopes, clearance and the verification gradient are the rails the system runs on — not a feature you switch on. Step outside the envelope and the action is refused.
Every delegation, decision and action writes to a complete, append-only audit lineage. When an auditor asks "who authorized this?", the answer is one query away — at every tier.
Built on CARE, PACT, EATP and CO — open standards anyone can read and audit. Your governance model isn't locked inside one vendor's black box.
Run Aegis where your data and policies live. Your information stays in your environment — see Security for the full posture.
If conditions change or a boundary is crossed, an agent's trust tightens instantly and automatically — the system never waits for a human to notice.
Aegis adapts to your departments, clearances and policies — generic by design, so it works for any organization without bending you to it.
The economics of governed autonomy, in plain language.
pseudo agent is inexpensive; a delegated agent doing the most work is where the value concentrates. Every quote is tailored to your deployment in conversation — we don't publish list prices.Tell us the function you'd start with and the autonomy you have in mind. We'll model a slice of your organization, show governance enforcing itself, and put together pricing tailored to you.