Fabric capacity from using Copilot in Smart Narrative Window

[sodra_sh]

Hi!

We are looking into building a report that utilizes the Copilot Preview in a Smart Narrative window. During testing it was noted that these reports use up quite a lot of Fabric capacity compared to our standard BI reports.

Is there any documentation, reference or examples that we can use to gauge how much X users or Y copilot calls will translate into Fabric capacity usage? Alternatively, are there any rules of thumb how one can convert the number of analyzed tokens in Smart Narrative to a ballpark cost?

Thanks!

[lbendlin]

Here's an interesting perspective (including costing)  Myths, Magic, and Copilot for Power BI — DATA GOBLINS (data-goblins.com)

[rajendraongole1 · Accepted solution]

Hi @sodra_sh - When utilizing the Copilot feature in Power BI, including the Smart Narrative window, it's important to note that the Fabric capacity usage is directly tied to the number of tokens processed by Copilot. Typically, 1,000 tokens correspond to about 750 words. This usage is tracked under your existing Power BI Premium capacity if you're subscribed, meaning that Copilot calls will draw from your Fabric capacity. Starting March 2024, Copilot billing is part of Power BI Premium, so it's essential to monitor your usage.

ref. links:

Fabric Copilot Pricing: An End-to-End example  | Microsoft Fabric Blog | Microsoft Fabric

Announcing Copilot in Fabric pricing  | Microsoft Fabric Blog | Microsoft Fabric

Copilot consumption - Microsoft Fabric | Microsoft Learn

 Solved: Need help to understand pricing $ for Copilot - Microsoft Fabric Community

Hope the above information helps , sharing links FYR