Book Image

Artificial Intelligence with Power BI

By : Mary-Jo Diepeveen
Book Image

Artificial Intelligence with Power BI

By: Mary-Jo Diepeveen

Overview of this book

The artificial intelligence (AI) capabilities in Power BI enable organizations to quickly and easily gain more intelligent insights from unstructured and structured data. This book will teach you how to make use of the many AI features available today in Power BI to quickly and easily enrich your data and gain better insights into patterns that can be found in your data. You’ll begin by understanding the benefits of AI and how it can be used in Power BI. Next, you’ll focus on exploring and preparing your data for building AI projects and then progress to using prominent AI features already available in Power BI, such as forecasting, anomaly detection, and Q&A. Later chapters will show you how to apply text analytics and computer vision within Power BI reports. This will help you create your own Q&A functionality in Power BI, which allows you to ask FAQs from another knowledge base and then integrate it with PowerApps. Toward the concluding chapters, you’ll be able to create and deploy AutoML models trained in Azure ML and consume them in Power Query Editor. After your models have been trained, you’ll work through principles such as privacy, fairness, and transparency to use AI responsibly. By the end of this book, you’ll have learned when and how to enrich your data with AI using the out-of-the-box AI capabilities in Power BI.
Table of Contents (18 chapters)
Part 1: AI Fundamentals
Part 2: Out-of-the-Box AI Features
Part 3: Create Your Own Models

Why should we use AI in Power BI?

The question of why we should use AI in Power BI is twofold. First of all, we may wonder why we should use AI to begin with, and secondly, we may wonder why we should use the features in Power BI. To answer the former, we need an understanding of what AI can do, which is covered in earlier sections. To answer the latter, we need to understand why AI is not being adopted yet by most organizations.

The problems with implementing AI

There is an undeniably large interest in anything AI related. Unfortunately, like many new technologies, everyone loves to talk about it, but only few actually do it. There are many reasons why the adoption of AI is lower than expected (McKinsey Survey on AI Adoption from 2018, accessed June 2021, The most obvious one seems to be the lack of skills. In previous sections, we discussed what AI is and how we can create models. We discussed how, to train a model, we need to choose an algorithm and that this requires data science knowledge, which is, among other things, a combination of mathematics and statistics. Consequently, a large reason of why companies do not use AI is because they do not have employees with the expertise in building these models and understanding the math behind it.

This is not the biggest roadblock. Many software vendors already recognized this problem some time ago and created tools and services directed at the citizen data scientist, democratizing the technology and making it available for everyone who wants to use it, regardless of whether you have a degree in data science. These easy-to-use tools should not replace any AI investments but do raise the question of why AI is not being used more.

The answer may be because people do not know about it. This has less to do with actual data scientists being hired but more with the rest of the organization. At the higher levels, leadership does not know how to form clear strategies or a practical vision around AI. At the lower levels, employees do not know how to use AI in their day-to-day work and even when AI insights are provided for them to work with, they often end up not trusting the machine as opposed to their own intuition. It seems that for both leadership as well as employees, the problem lies with an incomplete understanding of what AI can do and recognizing the possibilities and restrictions of AI.

Even when companies have recognized the potential, created a clear strategy on implementing AI, and hired the right people for it, they face issues. When you ask a data scientist what the most challenging part of their job is, they rarely mention training the model. The most common hurdle revolves around the data. Either for political or technical reasons, data scientists cannot get access to the right data. The data they need might not even exist. And even when they do get access to the right data, it is often not of good quality or of enough quantity to train a good model with.

And finally, to really implement AI into your business processes, you not only need data scientists, but you also need a whole team. You need data engineers to help you with extracting data from their sources, clean the data at scale, and offer it to the data scientists who can then train a model. After a model is trained, you need to make sure the insights are offered to the business in an intuitive way, for which you will need software engineers to integrate the model into the client applications or data analysts to visualize the insights from the models in your Power BI reports.

In other words, to create an end-to-end solution and implement AI in an enterprise environment, you need an interdisciplinary approach, and a collaboration between different departments, where, preferably, you also want to ensure that everyone has a basic understanding of what AI can do to build trust and enhance adoption.

Why AI in Power BI is the solution

Now that we understand the problem, we can get an idea of what could be the solution. Since there are many reasons there is a slow adoption of AI, here is an overview:

  • Lack of data science skills
  • Incomplete understanding of AI
  • Not enough, or not good enough, data
  • No collaboration between departments

Unfortunately, finding skilled data scientists is very challenging. The alternative can be to train your employees, which is something that will also help against an incomplete understanding of AI that is seen across different layers of organizations. To collect better or more data, we need to know what we are doing it for. Why should we invest in this and what will be the benefit? And to stimulate collaboration between departments, we need to create understanding and trust.

One tool that can help with all of these things is Power BI. Compared to data scientists, companies have significantly more data analysts who are familiar with working with data and who are either already working with Power BI or will easily adapt to working with it. That means that data analysts already know the importance of good data and have access to data. Using Power BI, we try to tell a story with the insights we generate from data, to help people make data-driven decisions. Data analysts know how to convey numbers into intuitive facts. They can help convey AI output into information that can be understood and trusted by anyone within and outside the organization. This can consequently also help with facilitating collaboration between departments as Power BI can already be used across different departments.

The only blocker is that those using Power BI are often not familiar with AI. They might not have the data science expertise, but they are the ones who can work very well with them. By combining AI with Power BI, we can educate others, to help the business create clearer AI strategies and to help end users gain trust over the model's output and how it can help them in their business processes.

That is why this book covers the different AI options in Power BI. They include low-hanging fruit to get started with today, to show the possibilities of AI. But Power BI can also integrate with sophisticated models that have been trained by data scientists. It is therefore a logical starting point to adopt AI at a larger scale within your organization.

Now that we understand why it is so beneficial to use AI in Power BI, let's have a look at our options.