Among the announcements Microsoft made in 2019 was AI Builder, a Power Apps tool that allows users to add AI-driven enhancements to apps in ways that Microsoft says will bring the value of artificial intelligence to an organization without the typical technical requirements.
With 20% of surveyed executives telling PwC they planned to implement AI enterprise-wide in 2019 at the same time as the recruitment time for tech roles is increasing, the timing for a solution that does not require coding skills to bring AI functionality into the workplace is good. However, unlocking the true productivity potential of AI Builder and similar tools requires both a realistic view of their capabilities and an understanding of the data used to drive these solutions.
It is the case that Microsoft’s AI Builder update to its Power Platform can allow non-programmers to add AI-driven functionality to applications, said Nikhil Koranne, assistant vice president of operations at software developer Chetu.
“Since AI, and automation in particular, is a hot topic right now for the tech and business fields, AI Builder offers an easy, more cost-effective solution for organizations looking to leverage AI without having to invest in their own costly platform or professional programmers,” Koranne said.
Users can choose from a variety of AI types that fit their business needs, connect data sources, tweak the models as needed, and train AI on business-specific formatting and data.
“Overall, the low-code capabilities of AI Builder make the integration of AI into platforms easier and much more inexpensive for businesses that can benefit from the automation and business intelligence gleaned from the AI,” Koranne said.
Benefits and Limitations
Tools such as Microsoft's AI Builder do hold potential for different types of enterprise operations. Angelo Frisina, the founder and CEO of app developer Sunlight Media LLC, pointed to inventory management as one example where one AI-driven feature – object detection – can have an impact. “G&J Pepsi, a Pepsi bottler and distributor in Cincinnati, Ohio, is using this technology to automatically identify and track their products,” he said.
But it’s important to remember that this functionality, while useful, also remains basic, Koranne said. It may fill the needs of some businesses, especially smaller ones, but will not fill the needs of all businesses. “For businesses looking to integrate AI solutions on a companywide level, it is recommended to look toward professional developers that can build and customize solutions to meet a business's needs and build a connected system that eliminates data silos and inefficiencies,” he said.
In fact, some of the benefits of AI Builder may come to those who do have a programming background. “It is the nature of any new, effective methodology to become reintroduced to the market in the style of building blocks,” said Alexander M. Kehoe, co-founder and operations director at Caveni. Kehoe brought up the example of WordPress, which built features that made it easier for anybody to produce a well-coded and well-designed website, but the people who benefited the most were already designers and developers, he said.
“In this same way, Microsoft’s new AI Builder and any other tools will do much the same for everyone already in the automation/AI space,” Kehoe said. “We can certainly expect more AI builders with new features and better ease of use to start popping up now as well.”
Tools such as AI Builder also have the potential to help automation become more accessible to smaller organizations, Kehoe said. “This heralds more of the advanced automation we have seen in larger businesses,” he said. “Small and medium businesses will now be able to automate many of the processes that tech giants like Microsoft and Google have already.”
However, the potential for Microsoft's AI Builder to reduce the staffing pressure for data scientists and AI experts isn’t quite there yet. Companies that are at the stage where they require such highly paid experts are further along than a tool such as AI Builder could take them, and those that can benefit from tools like AI Builder likely aren’t at that staffing stage.
While AI Builder might reduce the need for coders for a particular application, a firm, in-depth knowledge of data is still needed to optimize AI, Koranne said.
“I do, however, believe that with more ‘administrative’ work in app development being taken on by non-coders, it can free up specialists in data science and AI to invest more time into projects that can advance their company and drive innovation,” he said.
And some of the work of integrating AI into enterprise operations doesn’t involve the tech directly.
“Leaders also often think too narrowly about AI requirements,” Frisina said. “While cutting-edge technology and talent are certainly needed, it’s equally important to align a company’s culture, structure and ways of working to support broad AI adoption.”
Finally, it’s important to remember that AI-driven solutions aren’t something you implement once and then leave be. The tech is imperfect and evolving quickly, and what works now might not a year – or five years – from now.
“AI requires training, maintenance, testing and constant improvement,” Frisina said. For many organizations, that could mean a need for new staffers or increased pressure on CIOs, who have already seen their roles shift and expand considerably in recent years.