You know the agent works. Nobody can tell you what it takes to let a customer connect to it.
Get the agent out of the demo and in front of customers.
AI Agents & Platforms
Most organizations arrive here the same way. An agent was built and it works. Then a customer asked how sign-in is handled, or a launch date was set, or somebody asked what data it can actually reach. The build was the part the team knew how to do; everything after it is a different specialism.
What gets delivered
The agent is hosted where it stays up, sign-in runs through Entra, each data connection is scoped to what the agent needs, evaluation and telemetry record what it does, and the documentation a listing requires is written.
Which of these is on your list?
An agent that has to leave the demo
- You know it needs hosting that stays up. Right now it runs from one machine, or one account.
- You know a shared key will not survive a security review. Nobody has wired a real sign-in path.
- You know it has to be listed somewhere customers look. The documentation that requires has not been written.
A platform decision already taken
- You have chosen to build on Azure AI. Nobody in-house has stood up the model endpoints, the search layer and the identity around them.
- You know the retrieval layer decides whether answers are any good. Designing it keeps losing to shipping features.
Questions arriving from buyers, not from engineering
- You know enterprise buyers ask about security before price. The written answers do not exist yet.
- You know nobody can say what the agent returned last week, or whether it was right. Nothing is being measured.
What waiting keeps costing
An agent that only runs in a demo cannot be sold, so the work already spent building it returns nothing until the hosting, the sign-in and the documentation exist.
An AI platform built to make a demo work gets rebuilt when it meets real users, real data and a security review. The second build is the expensive one, and it lands on a date somebody has already promised.
Catalog and certification routes carry their own requirements and their own queues. Starting them late can add weeks, and those weeks land after the date a customer was given.
What made you look today?
Common reasons this becomes urgent
An enterprise customer asks how the agent authenticates and what data it can reach
Sign-in runs through Entra, every connection is scoped and listed with the permission it was granted, and the written security position exists before the next call.
A launch date has been given to a customer or a board
The server is hosted and reachable, the integration works on the platform the business runs, and the listing submission is ready to send.
How the work runs
The agent, its data connections and the target platform are reviewed, and the launch scope is agreed as a fixed piece of work. Then: The server is built and hosted, sign-in and connections are configured and tested, evaluation is switched on, and the documentation pack is prepared for submission.
Where this is priced
Each engagement below carries its own fixed scope. The price is agreed before any work begins, and the fixed proposal contains the delivery calendar and the completion date.
Scoped projects
A specific outcome whose complexity varies; price and calendar fixed after a short scope call.
Also delivered in this area, as custom projects
Work in this list is deliberately not packaged: the scope varies too much for one honest price. The route, the risks, the completion test and the fixed price are agreed before anything begins.
- Azure AI Platform Engineering
- AI Product Infrastructure Advisory
The finished state
What is true when it is done
A test user signs in through Entra, the agent returns only what that account could already reach, every connection is documented with its permission, and the listing submission is ready to send.
What you keep
Everything below is yours to keep whatever happens next, including handing it to your own team or another vendor.
- The hosted server and its configuration, written down
- The connection list, with the permission granted to each one
- The sign-in test results, per test account
- The evaluation scores against the agreed test set
- The documentation pack, ready to submit
For the full commercial shape of work in this area — scope, how the price is set, how you know it is finished — see a representative engagement: Public Agent & MCP Launch.
What moves the price
The fee is set by what is actually in the environment; headcount is only one of the inputs.
- How many APIs and data sources the agent connects to
- Whether it is going into a public catalog or staying inside your own tenant
- How much of the evaluation and telemetry has to be built rather than switched on
- How much of the security documentation already exists
Describe what happened
A reply comes within one business day with the most direct next step, or a clear answer that AZ Innovations is not the right fit.
Talk Through the Problem →