Connect the dots, close the loop

Accounts, customers, properties, and services lived in separate systems. Agents had to connect these in their heads for each request.

• Make the relationships clear. Each record shows its place in the request: which customer, which account, which property, and what services run there. Agents move between them without confusion.
• Let the request follow the agent. The current call or task, with its notes and open work, stays beside whatever record the agent opens. Agents never lose the thread.
• Start with the system, not the screens. I traced how records connected across the existing systems, then worked with domain experts and engineers to shape views agents needed and the team could build and support.

One request could set off several linked changes. In a move, every account, utility, and date had to be right at both ends.

• Show the change before it’s made. I made the account, address, utilities, and dates explicit in each request, so agents can verify every part.
• Offer the next step, not a dead end. In a move, arranging the stop offers to start service for the same customer. The agent picks up at the new address.
• Verify the outcome, not just the request. Testing showed agents needed to cross-check each service and its status. The flow keeps both in view, so agents can confirm what changed.

A call could end before the work did. Agents needed to follow every open item until it was done.

• Make follow-up visible. I made unfinished work part of the record. Open activities show workflows still in progress, and dated to-dos carry what’s left past the call.
• Make closing a deliberate choice. Finishing an activity cancels its pending workflows. A warning states how many before agents confirm, so ending the call doesn’t quietly drop the work.
• Keep enough context to return. Each to-do carries its due date, the related account, notes, and the call’s workflows and records. An agent can complete or forward it later without retracing the call.