How a Public Sector Institution Used AI to Classify Incoming Requests

The Challenge

The institution’s operations and support teams, along with its internal operations units, received work through inboxes and shared request channels. As the volume of incoming emails and requests grew, the initial work of reading each item, deciding what it was about and sending it to the right place began to slow down.

This pre-classification and routing was largely manual. Each email or request had to be opened and assessed before anyone could decide which category it fell into and who should handle it. Requests could wait before reaching the team or workflow responsible for them, and the time spent sorting incoming items affected the overall flow of operational work.

The Solution

With cognipeer Studio, our team configured an AI-assisted classification workflow for incoming emails and operational requests. Each item was analysed and categorised according to its content, so that the first sorting step no longer depended on reading each item by hand.

The solution then summarised the request and routed it to the appropriate workflow or to the relevant team. Requests that belonged in an established workflow were directed there, while others went to the team responsible for that kind of work. The teams receiving each request remained responsible for handling it, starting from a clearer picture of what it was about and why it had reached them.

Implementation

1. Defined request categories. We set out the categories that incoming emails and requests should be sorted into, based on their content and the teams and workflows that handle them.
2. Configured content analysis. The workflow analysed each incoming email or request to determine which category it belonged to.
3. Added summaries. Each item was summarised so that the receiving team could see its main points before opening the full message.
4. Mapped the routing. Categorised requests were directed to the appropriate workflow or the relevant team, which then took responsibility for handling them.

Business Impact

Incoming work reached the right place in a more organised way. Rather than sorting each email and request by hand, operations and support teams received items that were already categorised and summarised, and could move on to handling them with less preliminary effort. Fewer requests needed to be passed on after arriving in the wrong place. The path from a request arriving to the right team picking it up became clearer and more consistent.

• Faster initial assessment of incoming emails and operational requests.
• Less manual sorting across inboxes and shared requests for operations and support teams.
• More accurate routing to the right workflow or team.
• Better operational flow from intake through to handling.

Technology Behind the Solution