Usage Insights Agent · Use case
How can a utility use an AI agent to explain bills to customers on its own infrastructure?
A usage insights agent explains a customer's bill in plain language from that customer's own usage and tariff, and suggests ways to use less at peak times. A dispute, a payment difficulty or a change to the account goes to a member of staff. The agent runs on the utility's own infrastructure, so customers' usage data is not sent to a public AI service.
A bill nobody understands.
Then a call to the contact centre.
- ✗ 01Customers call because they cannot tell why a bill went up, and each call needs someone to read the usage.
- ✗ 02Usage data is personal, and a wrong explanation of a tariff causes a complaint.
Where it stops for a person
Disputes, hardship and account changes go to staff.A disputed bill, a customer in payment difficulty and any change to an account go to a member of staff. The subjects are set through guardrails.
It explains the bill.
Disputes go to staff.
Step · 01
Usage data and tariff documents are connected.
Step · 02
A signed-in customer asks about a bill.
Step · 03
The agent reads that customer's usage and tariff.
Step · 04
It explains what changed and suggests ways to use less at peak times.
Step · 05
A dispute or a request to change the account goes to staff.
Reads from · connectors
The agent connects to Snowflake or a Database for the customer's usage and bills, and to SharePoint for tariff documents.
Acts through · MCP servers
Nothing. It explains. It changes no bill, tariff or account.
What you set up
- Connect the usage data and the tariff documents.
- List the subjects that must go to staff.
- Test with past bill questions that have known answers.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Can it see other customers' data?
No. It is built to read only the records of the customer who is signed in.
Can it change a tariff or a bill?
No. It explains. Changes go to a member of staff.
Where is usage data processed?
On your own infrastructure, with a model you serve. Nothing is sent to a public AI service.