The recent launch of Copilot for Power BI marks a significant advancement in data analytics. With its general availability announced on May 20th, users with P1 or F64 SKUs can now leverage this powerful tool. However, its usage is contingent on the geographic region of your tenant. This blog post delves into the key features of Copilot, its benefits for different user personas, and best practices for report designers and data modellers.
Understanding Copilot’s availability
Copilot is now accessible to users with Power BI Premium Capacity (P1) or Fabric Capacities F64 and above. If your tenant is not in a region with Azure Open AI model availability, you can opt to turn on geographical data sharing to take advantage of Copilot for Power BI.
Maximising Copilot for your organisation
If organisations are hesitant to roll Copilot for Power BI to all users before understanding it’s capability, you can start with a smaller programme by restricting Copilot to a subset of users in your tenant settings. This approach allows for a controlled trial, ensuring that the organisation can adapt to the tool’s capabilities before a full-scale rollout.
End-user experience insight generation
End users benefit from Copilot’s ability to generate insights from reports. By asking Copilot for a summary, users receive key points to focus on, with references to the specific visuals used to derive the information. Report owners should be mindful of the data included in a report and advise end-users as such before taking Copilot at face value. For example, a report showing Actual v Target when only 2 months into the year could be interpreted by Copilot as being “behind target” when you display a full year’s target in your report visuals. In our testing, Copilot was not able to pick up on the context that we are comparing 2 month’s actuals with 1 year’s target. You could consider displaying a pro rata calculation of target in your report instead to allow Copilot to interpret this accurately.
Automated report generation for report builders
Report builders enjoy automated report generation, which not only suggests topics for insight but also creates comprehensive report pages with multiple visuals, offering a more in-depth analysis compared to traditional Q&A tools. Our testing has shown some fantastic results produced very quickly, and this could be key to getting report designers up and running with sleek and relevant reports. There are prompt limitations to be mindful of – you cannot specify filters to apply to your data (e.g. Last 30 Days) or ask for specific visuals. This can be easily added by report designers, so providing you see Copilot as a starting point and not a finished product, this will be a great value to organisations.
Enhanced DAX experience for data modellers
Data modellers experience an improved DAX query view, which accelerates the process of writing DAX queries. However, it’s crucial to be mindful of over-engineering and bloating the model with unnecessary calculated columns that can be produced depending on your prompt. Avoid prompts such as “Calculate Profit by Product”, as this may provide you with a calculation that summarises by Product, when all you need is the Profit DAX which you can then add as a measure and apply to dimensions as you wish. Our testing has found that prompts such as “Calculate Profit Last Year” are efficient, and will default to flags in Date Dimensions where available rather than using SAMEPERIODLASTYEAR or DATEDIFF.
What does this mean for my Power BI platform?
Many organisations find it challenging to strike the right balance between controls and innovation. Copilot is an exciting innovation, but deploying and using it with the right controls in place is key. Have you considered your workspace strategy that will allow users to save their Copilot produced reports with ease and not bloat your centralised workspaces? Worried about Copilot misinterpreting your tactical datasets that might not live up to best practice? Have you considered your Endorsement and Certification labelling to ensure that human-produced content which has passed QA can be easily recognised? These are all questions to ask in your BI function as you roll out Copilot and enjoy the benefits of quicker paths to insight.
Copilot for Power BI is a game-changer for data analytics, offering enhanced insights, automated reporting, and an improved DAX experience. With careful implementation and governance, organisations can enjoy the full potential of Copilot while maintaining control over their data models.
Join our Heads of BI & Analytics Megan Livadas and Colin Gresham as they discuss the launch of Copilot for Power BI in more detail below:
Interested in Power BI adoption and governance? Discover our expert Power BI Adoption solution or contact us to discuss your requirements today.