Ask Copilot helps you work with Power BI reports faster: it explains what a report shows, highlights what stands out in the data, and builds a chart or a table to support the answer. For instructions on how to open Copilot, see the article Ask Copilot into Power BI report.

Below are several situations in which Ask Copilot saves the most time.

1. Getting to know a new report

Instead of studying an unfamiliar report page by page, ask Copilot to explain it.

Open the report and ask: What is this report page about? Copilot explains the purpose of the page, its main indicator, and the metrics behind it.

In a few seconds, Copilot explained what the Cycle plan page tracks and summarized its key figures: 16,957 calls at 83% of the plan, 101,312 promotions at 229%, 11,577 covered targets at 88%, and an average frequency of 0.87 against a plan of 1.06. It also described how the page is structured — filtered to active users, with actual (AC) and planned (PL) values broken down by target group and category.

Instead of opening each visual and comparing the numbers manually, the user gets a ready summary of the whole page and can see at once which indicators are behind the plan.

Ask about a metric you do not recognize, for example: What does "Calls by target" mean and how is it calculated?

Copilot explained the Calls by target visual: how many calls were actually made to each target group or medical specialty compared with the plan. It also decoded the three columns — AC as the actual number of calls, PL as the planned number, and AC % as their ratio — using examples from the report itself, such as 2,463 calls to GPs/Physicians against a plan of 1,861.

This means users no longer need to ask a colleague or look through the documentation to understand an unfamiliar metric: the definition and the calculation logic appear right next to the data they are looking at.

Ask Prep a summary for my team to get a short text you can share with colleagues who are new to the report.

Copilot prepared a ready-to-share summary of the Cycle plan page: what the page tracks and how each key indicator stands against the plan — calls at 83%, promotions at 229%, coverage at 88%, and frequency at 0.87 against 1.06 planned. Each figure comes with a short verdict, such as below plan or well above plan.

Unlike the page description, this answer is written as a status update. Users can copy it into an email or a chat with the team instead of writing the summary and looking up every number themselves.

2. Getting insights from an existing report

Use Copilot to find out what stands out in the data instead of comparing the numbers manually.

Apply the filters you are interested in, then ask: Which criterion has the lowest score and why?

Copilot named the weakest criterion — Works effectively on multiple tasks simultaneously in the EFFICIENCY group, with an average score of 1.0 for two employees — and explained the reason behind it: the low ratings repeat across the last ten assessments rather than being a single bad evaluation.

This is the kind of answer that usually requires filtering several visuals and comparing them one by one. Here the user gets both the finding and its cause in one step, and can move straight to acting on it.

Ask about changes over time, for example: How has the overall score changed over the last months? Copilot gives a short conclusion and a chart that supports it.

Copilot described the trend across several years and built a chart to support it: calls and HCP coverage stay broadly stable with a slight upward trend in coverage, while promotion intensity drops from its 2023 and early 2024 levels and then levels off. It also warned that the last month's calls appear truncated, so the final point should not be read as a real decline.

The report itself shows only the current period, so answering this question would normally mean building a separate visual. Here the user gets the conclusion, the chart, and a caveat about the data quality in a single answer.

3. Viewing data in a way the report does not show

If the report has no visual for the breakdown you need, Copilot can build one from the same data.

Describe the breakdown you need, for example: Show the CRM activity metrics (calls, coverage, promotions) by representative and month as a table.

Copilot built a table that does not exist in the report: calls, HCP and HCO coverage, and promotions for each representative by month. It also added a conclusion drawn from that table — coverage and promotions per call stay almost constant for each representative, while call volumes fluctuate around those ratios, and some representatives keep a high reach while others maintain similar call volumes with structurally lower coverage.

Users no longer need to request a new visual from the analytics team to look at the data in a different way: the breakdown they need is built in the panel, while the report itself stays unchanged.

Copilot builds the table or chart in the panel, next to the report. The report itself is not changed.

Copilot turned the same request into a chart: monthly call volumes together with HCP and HCO coverage and promotions per call. The conclusion above it reads that call volumes stay broadly stable while coverage improves slightly and promotions per call fall from about 7 to 5–5.8 and then level off, with the last month standing out as an outlier because of low calls and missing coverage.

The chart appears only in the Copilot panel, so users can explore the data from a new angle without affecting the report or the way it looks for anyone else.

4. Comparing values quickly

Copilot can compare chains, regions, representatives, or periods without switching filters back and forth.

Ask a comparison question, for example: in the Cycle plan report: Compare plan completion for GPs/Physicians and Surgeons.

Copilot compared the two target groups across all three indicators at once: GPs/Physicians reached 132.3% of the call plan, 108.9% of coverage, and 132% of frequency, while Surgeons stayed at 73.6%, 73.4%, and 74%. The answer ends with a clear verdict — GPs/Physicians are outperforming Surgeons against the plan across all three KPIs.

Comparing two groups this way would normally mean applying filters twice and writing down the numbers in between. Here the user asks once and gets both sets of figures and the conclusion together.