Forum Discussion
Smart Narrative different between Power BI Desktop & Power BI Service
Hi,
I have created a smart narrative to summarize the data within my dashboard with copilot. I modify the prompt to copilot to generate the summary and insight with the smart narrative. And then i published the dashboard to Power BI service. But i found the smart narrative is different from my Power BI desktop. Anyone knows what happen and drive the different?
Thanks,
Nelson
Hi NelsonYP
This difference happens because Copilot runs in two different environments:
-
In Power BI Desktop, Copilot has access to your local model and the filters/slicers you’ve applied when you generate the smart narrative. The output is tied to the state of the data at that time.
-
In the Power BI Service, Copilot generates the smart narrative again based on the published model in the service, which may have differences such as:
-
Different data refresh timing (desktop uses your local dataset, service uses the refreshed cloud dataset).
-
Different filters or default slicer states applied when the narrative is generated.
-
The AI model version in the service may behave slightly differently than the one in Desktop, so the phrasing or emphasis of insights can change.
-
If you manually edited the narrative text in Desktop, those edits don’t always persist if the Service re-renders the Copilot-driven narrative.
-
-
4 Replies
- tayloramy
Super User
Hi NelsonYP,
What kind if difference are you noticing? Different wording is expected.
Copilot’s responses are nondeterministic, so even with the same prompt and data you can see different phrasing between Desktop and the Service. https://learn.microsoft.com/en-us/power-bi/create-reports/tutorial-copilot-power-bi-explore-dataIs this what you are experencing, or is there a bigger difference?
If you found this helpful, consider giving some Kudos. If I answered your question or solved your problem, please mark this as the solution.
v-karpurapudCommunity Support
Hi NelsonYP
Thank you for submitting your question to the Microsoft Fabric Community Forum, and thanks to tayloramy for offering helpful suggestions.
Could you let us know if the suggested solution resolved your issue?If you still need help, please share more details so we can assist you further.
Thank you.
- v-karpurapud
Hi NelsonYP
I wanted to check if you’ve had a chance to review the information provided. If you have any further questions, please let us know. Has your issue been resolved? If not, please share more details so we can assist you further.
Thank You. - Poojara_D12
Hi NelsonYP
This difference happens because Copilot runs in two different environments:
-
In Power BI Desktop, Copilot has access to your local model and the filters/slicers you’ve applied when you generate the smart narrative. The output is tied to the state of the data at that time.
-
In the Power BI Service, Copilot generates the smart narrative again based on the published model in the service, which may have differences such as:
-
Different data refresh timing (desktop uses your local dataset, service uses the refreshed cloud dataset).
-
Different filters or default slicer states applied when the narrative is generated.
-
The AI model version in the service may behave slightly differently than the one in Desktop, so the phrasing or emphasis of insights can change.
-
If you manually edited the narrative text in Desktop, those edits don’t always persist if the Service re-renders the Copilot-driven narrative.
-
-
