Agent Pilot
One working AI agent. Ten days. Fixed price. The fastest way to prove whether an agent can handle a specific workflow — before you commit to a full build.
Book a 25-min callProve it works before you scale it
You've identified a workflow that takes 5–20 hours per week of someone's time. It's probably document-heavy, repetitive, and the kind of thing your team does well but hates doing. You want to know if an AI agent can handle it reliably before you commit real budget to a full build.
The Agent Pilot answers that question. We build one production-ready agent in 10 days, connect it to one data source, integrate it with one system, and run 20 test cases before we deploy. You get working software in your environment — not a demo, not a prototype. If it works, you have evidence for the business case. If it doesn't, you've spent £6,500 to find out, not £25,000.
What you get
What you don't get
A multi-agent orchestration system. A full platform. A new way of working across your organisation. This is a focused, scoped pilot designed to prove value fast — or kill the idea before it wastes more budget. One use case. One agent. Done properly.
10 days from kick-off to live
Scoping & design
We align on the exact workflow the agent will handle, define inputs/outputs, map the data source, and sign off the integration plan. Everything scoped before a line is written.
Agent build
We build the agent, connect it to your data source, and wire up the first integration. Daily async updates. You can see progress in our shared workspace.
Evals & testing
We run 20 test cases against the agent — covering standard flows, edge cases, and failure modes. We fix anything that doesn't meet the bar before deployment.
Deployment & training
We deploy to your environment, run a 90-minute handover session with your team, and hand over documentation. You're live.
Post-launch support
14 days of support for any bugs, edge cases, or questions that surface in real use. We don't disappear at handover.
This is for you if…
- You have a specific workflow taking 5–20 hours/week of staff time
- You've decided to test AI agents but want a low-risk entry point
- You need working software in your environment to get stakeholder buy-in
- You're already a Microsoft 365 or Azure customer
This isn't for you if…
- You don't have a specific use case yet — start with a Discovery Workshop
- You need multiple agents across multiple workflows — see the Agent Build
- You need enterprise compliance or security architecture from scratch
Questions answered
What counts as 'one use case'?
One clearly scoped workflow — for example, 'answer HR policy questions from our SharePoint knowledge base' or 'triage inbound support emails and route to the right team'. If the scope is unclear at kickoff, we'll help you tighten it on Day 1.
Copilot Studio or Azure AI Foundry — which is right for us?
Copilot Studio is right if you're a Microsoft 365 shop and want the agent embedded in Teams or SharePoint. Azure AI Foundry is right if you need a more custom agent, multi-step reasoning, or integration with non-Microsoft systems. We'll recommend the right platform on the scoping call.
What if it doesn't work as expected?
That's what the 14-day post-launch support is for. If the agent isn't meeting the test cases we agreed, we fix it. The evals harness means 'meeting the bar' is defined upfront, not subjectively.
Can we use the agent code ourselves after?
Yes. Full ownership. You get the deployment, the documentation, and everything you need to maintain and extend it internally or with another developer.
What if we want to extend it after the pilot?
Most pilots turn into builds. The natural next step is an Agent Build (£18k–£28k) to scale the agent, add more integrations, or tackle a second use case. The pilot gives you the evidence to make that decision with confidence.
Ready to test it in 10 days?
Tell us the workflow. We'll confirm scope on a 25-minute call and give you a start date.
£6,500 fixed price · 10 days · Evals included