Power BI with Copilot...are you using it

[AndyDD_UK]

Not a desktop question but was interested if anyone is using Power BI with Copilot and how are you finding it ?

NB: I wasn't allowed to post on the Fabric community.

[Greg_Deckler · Accepted solution]

@AndyDD_UK Well, first, it is quite difficult to test Power BI with Copilot due to the extreme level of gatekeeping. Fabric trials do not provide the capability and so you must have a RESERVED F64 SKU or higher. In other words, even if you want to just kick the tires on it, it is going to cost you $60K USD. Not really feasible for a lot of folks out there.

Now, I am lucky enough to have access to an F64 reservation and I can tell you that it is pretty disappointing. I have found external tools such as ChatGPT and Claude far better than what Copilot provides and overall Copilot can't really do much that is even very interesting right now. I had to write some chapters on Fabric for some of my books and I wanted to include a chapter on Copilot as well but what Copilot could do, the actual reasons why you would use it, well, I had to drop the chapter because it was so bad and so limited in what it could do.

[v-bmanikante]

Hi @AndyDD_UK ,

We would like to follow up to see if the solution provided by the super user @Greg_Deckler  resolved your issue. Please let us know if you need any further assistance.

If our super user response resolved your issue, please mark it as "Accept as solution" and click "Yes" if you found it helpful.

Regards,

B Manikanteswara Reddy

[v-bmanikante]

Hi @AndyDD_UK ,

We would like to follow up to see if the solution provided by the super user @Greg_Deckler  resolved your issue. Please let us know if you need any further assistance.

If our super user response resolved your issue, please mark it as "Accept as solution" and click "Yes" if you found it helpful.

Regards,

B Manikanteswara Reddy

[v-bmanikante]

Hi @AndyDD_UK ,

Hi,

We would like to follow up to see if the solution provided by the super user @Greg_Deckler  resolved your issue. Please let us know if you need any further assistance.

If our super user response resolved your issue, please mark it as "Accept as solution" and click "Yes" if you found it helpful.

Regards,

B Manikanteswara Reddy

[Greg_Deckler · Accepted solution]

@AndyDD_UK Well, first, it is quite difficult to test Power BI with Copilot due to the extreme level of gatekeeping. Fabric trials do not provide the capability and so you must have a RESERVED F64 SKU or higher. In other words, even if you want to just kick the tires on it, it is going to cost you $60K USD. Not really feasible for a lot of folks out there.

Now, I am lucky enough to have access to an F64 reservation and I can tell you that it is pretty disappointing. I have found external tools such as ChatGPT and Claude far better than what Copilot provides and overall Copilot can't really do much that is even very interesting right now. I had to write some chapters on Fabric for some of my books and I wanted to include a chapter on Copilot as well but what Copilot could do, the actual reasons why you would use it, well, I had to drop the chapter because it was so bad and so limited in what it could do.

[AndyDD_UK]

Thanks Greg.

My understanding is that Power BI with copilot allow users to ask questions of the data and is more sophisticated than Q & A in that it allows follow up questions, more of a conversation, In additiona, copilot can be used to create reports/visuals

How have you used it ? What's good/bad from your experience ?

[Greg_Deckler · Accepted solution]

@AndyDD_UK It depends on where you are using it. In Power BI Desktop, you get prompts like "Create a new report page", "Suggest content for a new report page" and "Answer a question about the data". The content produced from things like "Create a new report page" are generally incredibly bad. You can ask things like "Create a new measure that does ..." but it can only handle really basic things and just provides the DAX code, you still have to create the measure yourself. It *can* carry on a basic "conversation" if you keep things relatively simple. It does remember the context of questions within a conversation. 

It's a similar story in the service. Basic functionality but most of the output is bad or limited. I just haven't found it as good or useful as the top tier models out there. It's just been "OK".