Call

Applied AI

AI governance and support for the AI tools you run

AI tools change without warning when a model updates or a connected system changes. Managed AI keeps every tool you run on a register, with an owner, reviewed access and regular checks on its output.

Request a proposal

Who it is for

Organisations with AI tools or automations already live, where nobody can say which tools the business pays for, what each one can reach, or who checks what it produces. Also anyone adding their first AI agent, because agreeing the controls for one tool takes far less work than untangling five.

What you get

An AI register that says what every tool and automation is for, who owns it, what data it can reach and what it is allowed to do. Its output is sampled against agreed checks, its access is reviewed, and the results appear in the monthly report you already receive.

The problem

Live AI tools drift. A model update or a change in a connected system alters what they do, and without an owner nobody notices until a client does.

Live AI needs an owner in the same way a server or a firewall does. Models are updated, connected systems change their fields and approval steps, and staff find new tools. Without a register and a review cycle, nobody can say which tools are in use or whether their output still meets the standard it was approved against.

Managed AI is that ownership, run by us alongside your service desk: a register of every tool and automation, shadow AI monitoring with the AI use policy kept current, regular sampling of output and drift checks on anything we built, a watch on the systems your AI depends on, access reviews, and a quarterly review that proposes what to do next and what to stop.

What Odyssey provides

  • An AI register of every tool and automation, its owner, its data and what it may do
  • Shadow AI monitoring, with the AI use policy kept current
  • Regular sampling of AI output against agreed checks, and drift checks on what we built
  • A watch on the systems your AI depends on, and fixes when something breaks
  • Access reviews, removing permissions a tool no longer needs
  • Usage and outcome reporting in the monthly report you already receive
  • A quarterly review that proposes what to do next, and what to stop
  • Licence management for AI tools, and support through the service desk

What is included

IncludedDetail
RegisterEvery AI tool and automation, its owner, the data it can reach and what it is allowed to do
Shadow AIMonitoring through DefensX and the AI use policy kept current
Output checksRegular samples against agreed quality checks, and drift checks on anything we built
DependenciesThe systems your AI relies on watched for changes, and fixed when they break it
AccessReviews that remove permissions a tool no longer needs
ReportingUsage and the agreed measures, in the monthly report
Quarterly reviewWhat to do next, and what to stop

Not included, or sold separately

Decisions made on AI output
Your staff stay responsible for what they do with it
Third-party model behaviour
We report it and work round it within our configuration
New builds and major changes
Scoped as Custom AI and Integrations

How it is delivered

  1. Build the register

    Every AI tool and automation in use, found through discovery and your own records, with an owner, a purpose and the data it may reach.

  2. Set the controls

    The AI use policy brought up to date, access trimmed to what each tool needs, and the checks agreed for its output.

  3. Watch and report

    Output sampled, dependencies watched, changes handled through the service desk, and the measures reported every month.

  4. Review each quarter

    What worked, what drifted, what to stop paying for, and what to do next, agreed with you.

Why trust Odyssey with this

  • ISO

    ISO 27001 and ISO 9001 certified by NQA

  • Our own staff work under the same shadow AI controls
  • We publish how we measure our own service desk

Questions people ask

Who owns an AI tool once it is live?

Someone has to, by name. AI tools drift as models update and connected systems change, so every live tool needs an owner, a review cycle and a way to switch it off.

What is Microsoft Agent 365?

Microsoft describes Agent 365 as a control plane for AI agents: one place to see which agents exist, what they can access and where they are in their lifecycle. It matters once a business has more than one agent, or agents it did not build itself.

What is an AI register?

A list of every AI tool and automation the business uses, with its purpose, its owner, the data it can reach and what it is allowed to do. It is the starting point for every review, and the first thing to fall out of date when nobody maintains it.

What do you report, and how often?

Usage of each tool and the measures agreed at the start, in the monthly report you already receive from us, with a fuller review every quarter.

Will the AI output always be right?

No tool is right every time. We sample output against agreed checks and act when it drifts, but your staff stay responsible for the decisions they make with it.

Can governance start with a single tool?

Yes, and it is the easiest point to start. Agreeing the controls for one tool takes a short conversation, and the register grows as the business adds more.

Ask about managed AI

Send us your question and we will come back to you. We call back within the hour in office hours.

Rather talk? Call 01642 661888. Already a customer with a fault? Call the service desk on 01642 661800.

So we can call you back

Fields marked * are required. We use your details only to reply to this enquiry (see our privacy policy).

Talk to a human about managed AI

Call and we answer within three rings. Or tell us what you need and we will call you back at any time.

01642 661888

Request a proposal Request a proposal