We’re excited to bring back Transform 2022 in person on July 19 and virtually from July 20-28. Join leaders in AI and data for in-depth discussions and exciting networking opportunities. Register today!
Artificial intelligence (AI) is demonstrating its ability to drive the growth of digital and non-digital native businesses. According to Deloitte, companies across all industries are using AI to create business value. From streamlining data analysis to improving customer experience, AI offers several benefits to businesses.
When AI is integrated into an organization’s core product or service and business processes, it is most beneficial. Despite the growing popularity of AI, many companies still struggle to use AI and ML on a larger scale. During a panel discussion at VentureBeat’s Transform 2022 virtual conference, Chris D’Agostino, Global CTO of Databricks, Patrick Baginski, Senior Director of Data Science and Analytics at McDonald’s, and Errol Koolmeister , AI and Data Advisor at The AI Framework, discussed how their companies are using AI and ML to create smarter customer experiences.
Implement AI and ML at scale
There is a growing interest in AI, its subfields, and related disciplines like machine learning (ML) and data science due to how AI is transforming every industry and business function. According to a recent McKinsey survey, 56% of organizations use AI in at least one business function.
Whether it’s a digital native or non-digital business, Baginski said it’s important to always think first about the value that can be delivered by AI and ML projects. According to Koolmeister, a 2019 MIT Sloan assessment showed how companies struggled as they persisted in trying to get their business off the ground, noting that AI’s return on investment was meager. Koolmeister also cited recent research by Thomas Davenport and NewVantage Partners that shows the market has changed — 26% of the world’s largest companies had AI in large-scale production while 92% are currently investing in technology.
“I think most companies are making efforts to implement AI in their organizations,” Koolmeister said. “There are a few things that are clear and distinct, one of them being internal capacity building to be able to provide big companies with AI. You can’t start with a decentralized organization, you have to build momentum First you need to create your first use cases and then build maturity as you roll things out in the organization so it needs to be learned by value and there needs to be clear proof points early on to adjust or actually motivate the levels of investment needed to transform large legacy businesses.
The Pitfalls of Trying to Create a Robust AI/ML Environment
Today’s AI is largely centralized and can only be owned by one entity. This is a significant hurdle for AI, according to Baginski, who noted that companies establish best practices, standard operating procedures and common platforms for the 80% of work done by analysts. , data scientists and data engineers. However, he affirmed that these activities should be seen as a collective enterprise that promotes remarkable development.
“I think one of the big challenges is forcing centralization,” Baginski said. “I think there’s a reason to say that you establish best practices, common platforms, and common processes for 80% of the work that an analyst, data scientist, or data engineer does, but you have to really see this as more of a community effort and your success in developing these guidelines depends on the company and business units adopting them, so forcing centralization is usually very detrimental to this effort.
Baginski also pointed to another challenge: moving from the generalist data science team that handles all of the machine learning, data science, measurement, analytics, pipeline building, etc., to several different roles. , more specialized, each playing a role in the big picture of building a good solution.
“The other challenge is that the devil is often in the details, isn’t it? So I think we’ve moved a bit away from the generalist data science team that’s just going to handle all the learning automation, all of the data science, all of the measurement, all of the analytics, all of the pipeline building and everything, to have several different more specialized roles, each playing a part in the bigger picture of developing a good solution,” Baginski said.
Baginski also noted that a typical challenge he’s seen is that a company needs to be very clear from the start, on a few project priorities or use cases that make sense for a team to start and that can be used to then essentially derive the adoption of these guidelines in the business units. He added that these use cases need to be properly reviewed by experts to determine how applicable they are to the idea of ML and AI, how well they serve that, and what value they will bring.
However, D’Agostino stressed the importance of building a team to solve the aforementioned issues.
“You won’t find a unicorn that will magically solve all of these problems. There really is a collaborative effort. Business stakeholders are key enablers to get things done. They understand what use cases need to be piloted within the company,” D’Agostino said.
Baginski said, “In a lot of companies, if you’re serious and you’re in a management C suite, you have to provide training or support without scaling so they can actually drive. So there is an educational aspect to being successful with these things.
Koolmeister added that the constant education of workers is absolutely essential, especially if it is in a large company that is highly distributed in many different countries.
Don’t miss the full discussion of what lessons McDonald’s, Databricks and The AI Framework have learned from implementing and scaling large AI initiatives to drive business value and smarter customer experiences.
VentureBeat’s mission is to be a digital public square for technical decision makers to learn about transformative enterprise technology and conduct transactions. Learn more about membership.