What you can ask

Once Rise is connected, you work entirely in plain language. The assistant figures out which report to use, what to filter and group by, and how to run it — you never touch report names, column keys, or query syntax. This page shows what to ask for and how the assistant handles it.

Example prompts

Quick answers

• "What's our current headcount?"
• "How many employees are in the Engineering department?"
• "What's the average salary across the company? And just in Sales?"
• "Who was hired in the last 30 days?"

Lists and lookups

• "List employees with an April birthday and their compensation."
• "Show me everyone terminated this year."
• "Which employees are in the Toronto office on the salaried pay type?"
• "Pull up Avery Tran's record."

Breakdowns and comparisons

• "Break down headcount by department."
• "Average tenure by location."
• "Compare total compensation between Engineering and Sales."

Build and save

• "Save April birthdays with compensation as a report called April Comp Review."
• "Add a Vancouver filter to my Headcount by Team report."
• "Open the headcount report in Rise."

Discovery

• "What reports do I have access to?"
• "What can I filter the payroll report by?"

What the assistant can do

Under the hood the assistant has a focused toolkit for working with Rise reporting. You don't call these directly — they're listed here so you (and anyone evaluating the integration) know exactly what's exposed.

Capability	What it does
Find a report	Browses the standard reports available to you and your own saved reports, and picks the one that matches your question.
Inspect a report	Reads a report's columns, types, and the valid values you can filter on — so it filters by real values, not guesses.
Run a report	Executes a query and returns the rows, with filtering, sorting, grouping, and totals/averages computed server-side.
Analyze results	Summarizes a result set — counts, distinct values, ranges, and top values per column.
Work with formulas	Looks up the available formula functions and validates a formula before using it, for derived columns like "birthday month" or "tenure in years".
Save & update reports	Saves a new custom report or edits one of yours — only if you've granted write access.
Link into Rise	Produces a deep link that opens the exact filtered report in the Rise web app.
Identify the session	Reports who it's signed in as and which build it's talking to, for troubleshooting.

Scopes

When you connect, you approve a set of OAuth scopes that bound what the assistant can do. Most assistants show these on the consent screen.

Scope	Lets the assistant	Needed for
reports:read	List, inspect, and run reports	Every question above that reads data
reports:write	Create and update your saved/custom reports	"Save as a report", "add a filter to my report"
user:read	Read your basic Rise profile	Identifying the session
user:write	Update your basic Rise profile	Rarely needed for reporting

For most reporting use, reports:read alone is enough; add reports:write only if you want the assistant to save reports for you. Granting fewer scopes never breaks read access — the assistant just can't perform the actions it didn't get. See the full scope reference.

Saved reports and deep links

The assistant treats your saved reports as first-class. Ask for one by name ("open my Q2 Comp Review") and it opens that exact configuration rather than rebuilding it. When you ask it to "open in Rise", it returns a link that lands you on the filtered report in the web app, where you can tweak, export, or share it with the full Rise UI.

Each turn produces one "Open report" link. If a question would need two different reports, the assistant will open one and offer the other as a follow-up, or ask you to pick.

Getting good results

• Be specific about thresholds and time ranges. "High earners" or "recent hires" are ambiguous — say "over $150k" or "hired since April 1". A vague ask gets a clarifying question, not a wrong answer.
• Ask for a breakdown when you want a total per group. "Average salary by department" returns one row per department, computed server-side — far faster and more complete than listing everyone.
• Expect a sample on big lists. A plain list caps at 100 rows for display. Ask the assistant to narrow, summarize, or save and open to work with the whole set.
• Let it look things up. If it asks which report or filter value you mean, that's it confirming against your real data rather than guessing.

Next steps

• New here? Start with the overview for the connection model and privacy details.
• Ready to connect? Pick your assistant: Claude, ChatGPT, Gemini, or Microsoft 365.