Build around the questions people ask now
Pull together a few common reasons people email or call your team. They might want to submit a request, upload a document, check who is handling an issue, or see what they need to do next. Ask a partner to walk through one of those situations and notice what language they use. Those words can become the labels in your portal.
Start with one request type and give users an easy way to see what they sent. A status such as ‘Waiting for your document’ tells them more than ‘Pending.’ Give each status an owner who knows when it should change and what message the user sees.
Give AI a specific job
An AI assistant could help someone find the right form, turn a long request into a short staff summary, or draft a response for staff to edit. Choose the job that solves a problem you can observe. If users cannot tell which form to choose, improve the labels and test those before deciding whether to add chat.
When AI answers questions, choose the approved material it can draw on and give users a way to reach a person. Test what happens when the policy is unclear, a document is outdated, or someone asks about a situation the portal does not handle. Staff need a clear path to take over.
Test what each person can see and do
Write down who uses the portal and what each role needs. A partner may need access to their own submissions. A staff reviewer may need a queue across partner accounts. A program lead may only need totals. Test these views using sample accounts before inviting people to use the portal. Include a test where a user changes organizations or loses access, so staff know how to update that account.
Ask a small group of users to submit a request and find its next step without help. Also try a rejected upload, a duplicate request, and a missed notification. A successful test shows that people understand where they stand and staff can resolve mistakes without searching across inboxes.
WORK WITH LESS DESK