Ambiguous requests. Requirements had to be gathered in detail before planning, with repeated back-and-forth on what was meant.
Requirements agent built in Microsoft Copilot Studio; AI-guided questions gather detail at intake.
The team starts from a clearer picture of each request before planning and execution.
- Request
- AI-guided questions
- Detailed requirements
- Structured intake
- Delivery planning
- Execution readiness
A delivery team in a global in-house centre handles requests from brands across a large organisation. Each request needs to be understood before it can be planned.
Requirements have to be gathered in detail before a team can plan against them, and gathering them well is repeatable work.
Every unclear request costs the team time in back-and-forth before work can start.
I designed and built the agent in Microsoft Copilot Studio, working with Microsoft Copilot.
The gap was at intake. Client servicing couldn't always capture the complete requirement or know the right questions to ask the stakeholder, so the delivery team filled the gaps with assumptions and the work went back and forth. Fixing intake fixes the problem at its source.
The agent asks the questions on the team's behalf, so client servicing doesn't need to know every question each team would ask. It keeps asking until the request is clear, then produces a short brief in plain language that the delivery team can understand and act on.
Requirements are now gathered in detail through the agent, so the team starts from a clearer picture of each request.
I introduced it to the team with a live demo and put it to work on incoming requests.
The team understands requests more easily, which means they can get on with delivering instead of going back and forth on what was meant. Less back-and-forth, fewer assumptions and time saved at intake.
The cheapest place to create clarity is at the very start of the work.