Power BI AI features include AI visuals like decomposition tree, smart narratives, Q&A and Key influencers. Other AI features inside Power BI charts are automatic anomaly detection, explaining increase in line chart and auto-clustering in scatter chart.
There are many more AI features in Microsoft Fabric including the all-powerful Microsoft co-pilot. If you want to read about those check out our Power BI vs Microsoft Fabric article. You might also be interested in our guide to Linear Regression in Power BI.
In this article we will aim to demonstrate these features and also give step-by-step instructions on how to access them. Speak to a Power BI consultancy if you want to get help to implement these features effectively in your organization.
Let’s explore the various Power BI AI features and visuals that you can find outside of Microsoft Fabric.
The explain increase/decrease feature is an advanced AI-driven capability that empowers users to analyze and understand the reasons behind trends or variations in their data. This feature is particularly useful in line charts to identify contributing factors for spikes or dips in performance metrics, allowing businesses to make informed, data-driven decisions.
Here’s a detailed explanation of how this feature works, supported by an example using sales data.
Open your report and create a line chart. Right-click the data point and choose Analyze > Explain the increase or Analyze > Explain the decrease.
Power BI examines related dimensions such as product categories, channels, and customer segments to determine the factors responsible for the change. It then provides these insights through visual explanations. In this case, the analysis reveals that the increase in revenue is primarily attributed to the Wholesale channel.
This AI-generated explanation also highlights how the distribution is influenced by filters, such as the Retail and Wholesale channels, which together contribute to 50% of the total revenue.
Anomaly detection in Power BI is an advanced analytics feature that highlights data points that deviate significantly from the expected pattern or trend in a line chart. These anomalies are accompanied by explanations powered by machine learning, helping users understand the potential reasons behind the outliers.
In Power BI Desktop, drag your data fields to create a line chart. Ensure the chart has a measure (e.g., sales, profit) on the Y-axis and a time-based field or category on the X-axis.
Select the line chart and go to the “Analytics” pane. Add the “Find anomalies” option and Power BI will automatically scan the data and highlight anomalies with markers.
Adjust the “Sensitivity” slider to control the anomaly detection strictness. A higher sensitivity identifies more anomalies, while a lower sensitivity focuses on major outliers. You can also choose specific data fields to include in the explanations for more relevant insights.
Click on an anomaly marker to view the possible reasons for the deviation. Power BI generates explanations using the underlying data fields.
We have several tips on effectively using this feature in your Power BI reports:
Auto-clustering is a Power BI feature that assigns data points to groups or clusters based on their characteristics. When applied to a scatter chart in Power BI, the algorithm identifies and assigns each data point to a cluster, visually separating the groups with unique markers or colors. This feature can reveal hidden patterns, such as customer segmentation, product performance categories, or geographic groupings.
Before you use this feature ensure your dataset is clean and includes numerical or categorical fields suitable for scatter chart visualization. Select scatter chart from the visuals:
Drag and drop your fields into the X-axis and Y-axis of the scatter chart.We added date in the values and revenue and order in the x and y axis respectively.
Select the visual and Click on the “More Options” (ellipsis) in the top-right corner of the visual. Select “Automatically Find Clusters. Power BI applies machine learning to identify and group data points into clusters.
Adjust the number of clusters based on your analysis requirements. You can increase or decrease the number of clusters to refine the groupings. Rename clusters for better clarity and understanding.
Review the visualized clusters to gain insights into relationships or trends within your data. Use cross-filtering to explore cluster behavior in relation to other report elements.
Power BI has created 5 different clusters based on data. We can further format the clusters to show better visualizations.
Here are a few tips to help you make the most out of this feature:
The Q&A visual enables users to type questions about their data directly into a report. Power BI interprets these questions and generates visualizations, such as charts or graphs, to provide answers. This feature is powered by a natural language processing engine, making it intuitive and user-friendly. Here’s a detailed overview of the Q&A visual, along with an example:
You can create the Q&A visual in Power BI Desktop, double-click anywhere on the report canvas or select the Q&A visual from the visualization pane.
Type your query into the Q&A box. For example, “Sale by channel.”
Power BI will create a visual tailored to the requirements, and in this example, it generates a table displaying sales by channel.
One challenge with the Q&A visual is ambiguity. It could be that many of your report users are interested to see the list of biggest clients. However, they can ask this question in a number of different ways: biggest customers, consumers, shoppers, etc.
Luckily, Power BI report admins can see all the requests typed into the Q&A visual and can fine-tune the performance of Q&A visual. You can enhance each table with additional synonyms by clicking on the “Add Synonym” button within the visual.
You can further customize the visual by changing the chart type or adding filters to focus on specific regions or product categories.
The Decomposition Tree is an AI-powered Power BI visual that enables users to analyze data hierarchically. It aggregates data and allows users to drill down into dimensions in any order. This flexibility makes it ideal for scenarios where users want to explore data dynamically and uncover patterns or anomalies.
You can create a decomposition tree visual In Power BI Desktop by selecting the Decomposition Tree icon from the Visualizations pane.
Drag a measure (e.g., sales, profit) into the “Analyze” field. You can then add dimensions (e.g., region, product category) to the “Explain By” field like we added channel and customers to it.
Click on the “+” icon to drill down into dimensions and perform a hierarchical analysis of the data. By default, high and low values will be displayed, while all other dimensions actively used in the “Explain by” tabs will also be shown
Select any dimension to view the revenue segmentation. For instance, choosing a channel displays the revenue for each channel, and within a specific channel, Power BI further shows the revenue generated by individual customers, as illustrated below.
.
The selection of dimensions is user-driven, allowing them to analyze data in any preferred order. Additionally, AI splits can be utilized to let Power BI recommend the next level of analysis
The Smart Narratives visual is designed to generate text-based summaries of your visuals and reports. It uses natural language generation (NLG) to analyze data and create narratives that highlight important trends, outliers, and key metrics.
This feature is ideal for users who want to add context to their reports without manually writing detailed explanations.
You can can create this visual In Power BI Desktop by clicking on the Smart Narratives icon in the Visualizations pane. Currently, Power BI is integrating Copilot with this visual, giving you the option to choose between Copilot or a custom setup. For now, we’ll proceed with the custom option
To summarize an entire page, click on the report canvas and add the Smart Narratives visual. This will automatically create a summary of all the visuals on the page. To summarize a specific visual, simply right-click on it and select “Summarize.” It generated the following summary for the entire page.
Utilize the text box commands to modify the generated content. You can customize the font, color, and formatting according to your preferences.
The key influencers visual allows users to analyze their data to understand the factors that influence a selected metric. For example, it can answer questions like “What drives customer satisfaction?” or “What factors contribute most to sales growth?” By leveraging statistical techniques, this visual identifies patterns and relationships in the data.
You can create this visual in Power BI Desktop by selecting the Key Influencers icon from the Visualizations pane.
Place the metric you wish to analyze (e.g., customer satisfaction score) into the “Analyze” field. Then, include dimensions or fields that might impact the metric (e.g., age, income, region) into the “Explain By” field. In this case, we’ve added the customer satisfaction score to the “Analyze” field, and region, delivery time, and loyalty program membership to the “Explain By” field
The visual presents a ranked list of factors affecting the metric, accompanied by visual explanations. For instance, it shows that as the average delivery time decreases, customer satisfaction improves. Additionally, filters can be applied to highlight specific aspects of your data and enhance the analysis further.
These insights enable the company to focus on improving delivery efficiency, tailoring services for rural customers, and optimizing order management processes.
Microsoft Power BI has evolved into a powerhouse for business analytics, and its integration of artificial intelligence (AI) capabilities has revolutionized how organizations interact with data. From automating insights to enabling natural language interactions, Power BI AI features empower users to uncover hidden patterns, predict trends, and make data-driven decisions faster.
