The task
Someone in finance, sales or product needs a number: revenue, costs, sales, or how a product is doing.
Before, that meant asking an analyst or ordering a dashboard. That could take weeks.
Finance, sales, product and operations
Answers from company data, without waiting for an analyst
People ask about the company’s data in plain words. The agent writes the queries and answers with a table.
Active customers by region
How did weekly active customers change by region, last month against this month?
Data warehouse List tables · customers, activity
Data warehouse Run query · column renamed, fixed and ran again
Code Count weekly averages
Weekly active customers grew in 3 regions out of 4. The South fell by 6%, mostly in the last 2 weeks.
RegionLast monthThis monthChange
North4,2104,480+6%
Central6,9307,150+3%
West3,1203,260+4%
South2,8702,700−6%
Message…
AgentData Analyst
ModelGPT 6.1 Sol
ReasoningAuto
Screens show sample data. The case is real use in the group that built Sigma.
How the agent does it
- An agent gets access to 26 private tables, so different people can ask about that data.
- A person asks in a chat, like asking a colleague.
- The agent reads the list of tables, writes a query and runs it. If the query fails, it fixes it and tries again.
- It does the maths in code and answers with a table or a chart.
- It uses the person’s own access. It can’t show them data they are not allowed to see.
What changed
- People get the answer in the chat. No analyst request, no new dashboard.
- Follow-up questions go in the same chat, and the agent carries on.
In numbers
Reach
34%of people who use agents asked data questions this way: 205 people
Last 30 days
Frequency
1,301conversations with data queries in 30 days
Last 30 days
Finishedcompany-wide
88%of agent runs across the company are completed
Last 30 days
Time
~6 minper conversation with data queries (median), including the person’s own pauses
An earlier 30-day period
Production data inside the group, read-only. A run is finished when it ends with the status “completed”.
More from our own use
Client teams, operations and planning
Reports that build themselves on a schedule
Agents build daily and weekly reports on their own and post them where the team works.
Read the case Technical supportA first look at every incident ticket
The agent picks up each new incident ticket, gathers the facts and posts the likely cause to the support chat.
Read the case Customer supportFirst line of customer support
In one of the group’s products, the agent takes new customer requests, looks up the account and brings in the teams it needs.
Read the case