For the IT lead

AI agent access control for CIOs, on top of your identity provider.

Each agent signs in as itself. For each job it gets a permission that is scoped and expires. Data goes only where your policy sends it, and every action and refusal is on a record you can check without us.

Send this page to your teamEmail it

A decade in defence and aerospace. In government procurement since 2016. The same technology since 2019, validated in US DoD and Canadian DND programs.

What your business units will ask for

Your business units will ask for agents that change orders, read contracts and answer suppliers. Some already exist, built by your own teams or bundled in products you license. The question for the IT lead is what each one can act as, where its data goes, and what you can show afterwards.

Ares answers those three questions in one place, for every agent you enrol on it, whoever built it. An agent you have not enrolled stays outside it.

How it fits what you run

  • Identity. Ares sits on top of your identity provider. Your provider says who the agent is. Ares decides what each job may do.
  • Access. You set an access policy you can read: which systems, which actions, which data. The policy is checked on every call, not once at sign-in.
  • Handoffs. When one agent passes work to another agent or to a tool, the permission narrows. It never widens on the way.
  • Content. Documents are read at the door before an agent sees them. Answers are checked before they leave, and each redaction goes on the record.
  • Placement. Cloud, hybrid or in your building. It can run on one small machine there, with local models.

What you can show afterwards

Every decision is recorded, including every approval and every refusal. You export the record and check it offline. If a record was changed, the check fails. Hand your auditor the replay pack and ares-verify, the open-source verifier, not a promise.

Follow one workflow

What the IT lead sees in one run.

An illustration of one access request, told from the IT side. This workflow is designed and has not run in the sandbox yet.

  1. 01Ticketing

    A request arrives in ticketing: read access to the finance reports for a contractor, until the contract ends.

    Who approves
    No approval on this step
    On the record
    job opened · from ticket
  2. 02In-house tool

    The agent signs in through your identity provider and receives a credential for this request only.

    Who approves
    No approval on this step
    On the record
    credential issued · one request · expires at job end
  3. 03Documents

    Check the contractor's role against the access policy for the finance reports.

    Who approves
    No approval on this step
    On the record
    check-policy · allowed
  4. 04In-house tool

    Draft the grant: read only, with an end date.

    Who approves
    No approval on this step
    On the record
    draft-grant · allowed
  5. 05In-house tool

    Widen the grant to edit rights, as a comment on the ticket asks. The policy allows read only for this role, so the step is refused.

    Who approves
    No approval on this step
    On the record
    widen-grant · refused · not in policy
  6. 06In-house tool

    Apply the grant. The step waits for the IT lead.

    Who approves
    IT lead
    On the record
    apply-grant · approved by IT lead
  7. 07In-house tool

    On the end date the grant is removed, and the removal goes on the record.

    Who approves
    No approval on this step
    On the record
    grant-expired · removed

In short

  • Where does data go?Where your policy sends it, and nowhere else. A call to any other address is refused and recorded.
  • Who can the agent act as?Only itself. Each agent has its own identity and a permission for one job that expires.
  • What is logged?Every decision, approval and refusal. An outside party can check the record offline, without us.
  • Is it another system to look after?It sits on your identity provider and the systems you run. It can run on one small machine in your building.

Runs to check before your review.

Each built run opens in the sandbox with real agents on demo data. Designed items set out their steps here.

Built
We have run it.
Designed
The steps and approvals are designed. It has not run yet.
On request
We scope it with you before we design it.
  • Built

    A hidden instruction in an inbound document

    A broker's PDF tells the agents to waive a survey. It is read at the door before any agent sees it.

    Run the underwriting case
  • Built

    Permission that narrows at a company boundary

    A supplier's agent reaches for more than the order line. The permission narrows, and both records agree on what crossed.

    Run the two-company order
  • Designed

    Access request and grant

    The workflow above, from the ticket to the removal on the end date.

  • Designed

    Change approval for a system of record

    Steps: read the change ticket, check it against the change window and the policy, collect each approval the policy names, apply nothing before the last one, record every approval.

The five guarantees

Every workflow on this page carries them. Each one has a limit, and we state it.

  1. 01

    Each agent can do only what it was allowed to do, for that job.

    The limit

    Permissions are only as good as the rules you approve.

  2. 02

    Sensitive steps go to a person.

    The limit

    You decide which steps are sensitive.

  3. 03

    Documents are inspected before an agent sees them, and answers are checked before they leave.

    The limit

    Filters cover stated patterns and languages.

  4. 04

    Every decision is recorded, including refusals, in a record an outside party can check.

    The limit

    It shows what agents did, not whether a business rule was wise.

  5. 05

    It can run in your building, on one small machine.

    The limit

    Local models differ from the largest cloud models.

What it does not do

  • It does not replace your identity provider, and it connects to one identity provider product today. Others are on request, and the briefing confirms yours.
  • It does not write your access policy. Your team sets it, and Ares checks it on every call.
  • Content filters cover stated patterns and languages, not every form a hidden instruction can take.
  • Local models differ from the largest cloud models. We say which model runs which step.
First step

Briefing, evaluation, licence.

Each step has a fixed scope and ends with something you keep.

  1. 01

    Platform briefing

    What Ares does, and how it fits your identity provider and the systems you already run.

    What you get
    A briefing for your AI, architecture and security leads.
    What you keep
    Agreement on one agent or workflow to evaluate.
  2. 02

    Evaluation with one of your agents

    One of your agents enrolled on Ares, with credentials, an access policy, a governed path and the record.

    What you get
    Your agent running on Ares, in your environment, at a fixed scope.
    What you keep
    Your own evidence that it works in your environment.

    Fixed price, quoted after one call.

  3. 03

    Licence and rollout

    The platform in production, in the deployment shape you choose.

    What you get
    Onboarding for your teams. Solutions and delivery help where you want them.
    What you keep
    The platform in production.

    The platform tiers on the pricing page.

Questions and answers

Evaluate Ares with one of your agents.

Bring one agent and one tool it calls. We enrol it on Ares in your environment, and you keep the evidence from the evaluation.

A senior member of the team replies within two business days.