Skip to main content

Announcing new AI Capabilities for Power BI to make AI Accessible for Everyone

Headshot of article author Arun Ulagaratchagan

At Microsoft, our mission is to empower every person and every organization on the planet to achieve more. Power BI makes it possible for every employee in an organization to make better decisions based on data with beautiful reports and dashboards.

With the massive volumes of data generated today about every aspect of a business finding insights from the data can be challenging. This is where AI can help. AI can aid in data exploration, comb through the data to automatically find patterns, help users understand what the data means, and predict future outcomes to help business drive results.

There are huge opportunities when workers across an organization can leverage AI for improving business outcomes. But data scientists, who are crucial to deploying AI solutions, don’t have the bandwidth to provide custom solutions to all users at an organization. Businesses need ways to surface the valuable work of data scientists and empower more users to leverage AI in easy and intuitive ways. 

Power BI has been a pioneer in applying AI through capabilities such as natural language, which enables users to get answers by asking questions in plain English, or Quick Insights, which automatically finds patterns in data.

Today, we’re taking a major step forward in bringing AI to business intelligence and announcing several new AI features in Power BI that are available in preview.

  • Users can now get capabilities such as image recognition and text analytics directly in Power BI. 

  • Key driver analysis helps users understand what influences key business metrics. 

  • Users can create machine learning models directly in Power BI using automated machine learning.

  • Users now have seamless integration of Azure Machine Learning within Power BI.

All these new AI capabilities—pioneered in Azure and now available in Power BI—require no code. This enables all Power BI users to discover hidden, actionable insights in their data and drive better business outcomes with easy-to-use AI.   

Azure Cognitive Services capabilities in Power BI

Azure Cognitive Services are sophisticated pre-trained machine learning models that can extract insights from data. We are bringing Azure Cognitive Service capabilities into Power BI to provide powerful ways to extract information from a variety sources like documents, images, and social media feeds. These algorithms can identify named entities such as organizations, people, and locations. They can recognize objects in images, detect language, identify key phrases, and determine positive or negative sentiment.

Imagine you’re a business analyst for a hotel chain and want to evaluate what guests are saying in their online reviews. With Azure Cognitive Services in Power BI, you can easily analyze thousands of online reviews, understand what your guests are happy or unhappy about, and pinpoint areas of improvement. In the example below, the user can understand that a specific hotel has issues with air conditioning, causing customer dissatisfaction and negative reviews. Azure Cognitive Services capabilities in Power BI can surface this insight automatically, so the hotel can take action.   

HotelReviews

New Key Driver Analysis feature in Power BI

Every organization has metrics or key performance indicators that measure business success. How do businesses determine what impacts those KPIs? What causes these KPIs to go up, down, or stay the same? The most obvious business drivers are typically easy to spot, but it doesn’t take long before things get complex. What drives business outcomes is often subtle and depends on any combination of circumstances. 

Key driver analysis helps you understand what drives an outcome. It reasons over your data, ranks those things that matter, and surfaces those key drivers. For example, consider a student’s plans to attend college as a KPI. There are different factors that impact whether kids plan to enroll in college. Key driver analysis automatically surfaces those things that matter the most. Below, you see that parental encouragement has significant positive impact on a student’s plans.

Build Your Own Machine Learning Models in Power BI 

In Power BI, business analysts will now be able to build their own machine learning models without writing a single line of code. We’re using the automated machine learning feature in Azure Machine Learning, but instead of targeting developers or data scientists, we’ve simplified it and made it broadly accessible for common use cases. This means that when an analyst builds a machine learning model in Power BI, it does all the heavy lifting by selecting the best algorithm and features with just a few clicks.

As an example, a business analyst could leverage the automated machine learning technology to quickly and easily build a model to predict how likely an open sales opportunity is to be won. This could help a sales manager prioritize which high value opportunities to focus on and how likely they are to meet their target.

Integrate Your Azure Machine Learning Models with Power BI

Advanced machine learning requires specialized data science tools. Azure Machine Learning is a platform where data scientists develop machine learning models to take on complex business challenges. Azure ML models built by data scientists can now be easily shared with business analysts. Power BI works behind the scenes to discover the models to which each user has access and automatically creates a point and click user interface to invoke them. This makes collaboration among business analysts and data scientists easier and faster than ever before. 

Sign up for the preview to try new AI capabilities in Power BI

Power BI is building on years of work in Microsoft Research and Azure, and bringing these capabilities to any business user, regardless of their coding skill. Complex tasks that typically require technical know-how — key phrase extraction, sentiment analysis, understanding drivers, creating machine learning models—will now be possible with just a few clicks and without code. This empowers everyone in an organization to harness the power of AI to make better decisions. 

Sign up for the preview here.