The project ends. The agent keeps talking to your customers.
After go-live, the questions nobody was asked to answer are still open. Who is accountable when the agent is wrong? Who may edit its instructions, and does anyone check the edit first? How would you switch it off? How do you know it is working? That gap is where agents drift, lose trust, or get turned off.
One written report, with the fixes in order
In writing, with no meetings
- Two people answer a questionnaireThe person responsible for the service and your Salesforce admin, separately. About 30 minutes each.
- Your admin runs two short commandsAn export of the agent's configuration and a list of its versions. We never log in to your org, and no customer records are needed.
- You receive the checkupThe written report and a recorded walkthrough, within 10 working days. Questions are answered in writing.
See what a finished checkup looks like
The sample is for a fictional company. Three of its findings are modeled on what a real export showed in a test org running Salesforce's default service agent.
From the side that hands the system over
Fixed, so you know before you start
In development. XP Checkup is not open for orders yet. If you have an agent live or in rollout and want to be among the first, write to checkup@xenessa.ai.