Want to Leverage AI in Your Business? Read These 6 Facts

Fintech Icon

Want to Leverage AI in Your Business? Read These 6 Facts

Profitect recently hosted a private dinner on the future of artificial intelligence (AI) in data analytics. The keynote speaker of the night was Forrester Research VP and Principal Analyst Mike Gualtieri. Gualtieri is considered a premier expert on the intersections of business, AI, and innovation and has done extensive research into such fields as AI, data analytics, and machine learning. Here are some of my favorite takeaways from his presentation.

  1. AI is not necessarily here to replace humans

In response to an audience question regarding this misconception, Gualtieri provided a historical example of what happens to jobs when technology advances. He cited the introduction of word-processing software, which did indeed cost many typists their jobs, since those who had once relied on them now had word-processing software right at their desks. However, while AI would likely cost some jobs, Gualtieri stressed that many of these jobs and others would simply transform. New jobs would appear and people will be paid more through the introduction of AI-powered solutions that add more value than ever before. For revenue-minded retailers, this is a welcome breath of fresh air.

  1. With data, less can be more

Gualtieri was asked by the audience how much data one needs in order to harness AI. Gualtieri’s response? Yes.

“It’s true that you need an awful lot of data in some cases; it’s truer that you need the right data,” he said. The exact amount of raw data to analyze is irrelevant in most cases — it’s the variables within said data that AI needs in order to work. According to Gualtieri, a billion lines of data would mean nothing without those variables.

Gualtieri also emphasized that retailers have no need to worry about this, since their transactional data is rich in these critical variables. “If you’re in a retail situation that’s ideal because you have [data from] transactions,” he said. Transactional data is naturally ideal for AI processing because it’s a direct window into the most valuable parts of the customer relationship: when the money changes hands, the basket size, the number of items being purchased, time between transactions, traffic information, loyalty information, and more.

  1. There’s more than one type of AI

When people think of AI, they often think of what Gualtieri called “sci-fi stuff,” such as artificial brains, intelligent robots, and what we’ve see in movies like Star Wars and the Terminator series, to name a few. While not entirely inaccurate, this is only one of two types of AI. The aforementioned examples fall under the umbrella of “pure AI” — essentially building exact duplicates of human intelligence in the form of a machine. The more functional type (and the one that retailers need to learn to harness), is “pragmatic AI” — the automation of business processes to drive value and help humans make more informed and optimal decisions.

  1. AI is not its own technology

While many people believe AI to be its own self-powered technology, this is a myth, Gualtieri noted to the packed crowd. Rather AI is an umbrella term, a technology powered by one or more building-block technologies. For example, the AI used by Profitect is “pragmatic AI,” powered by two building blocks: machine learning and prescriptive analytics (delivering the analysis results plus corrective actions needed to address them and therefore optimize the outcome).

  1. AI is not necessarily better than humans

Humans and AI have a complicated relationship. Machines certainly have an edge over human intelligence in terms of capacities — but this isn’t because machines are smarter. They are simply much faster. In addition, they only do exactly what they’ve been “told” (programmed) to do, making blind spots a very real possibility. Humans, on the other hand, can make informed decisions too — but our capacity limits drive us to take shortcuts. We form almost a symbiosis with machines by adding a feedback loop, according to Gualtieri. When a machine analyzes and interprets data with more than 1,000 factors and draws a conclusion, the best efficacy comes from having a human operator to verify or invalidate said conclusion. The machine can “learn” and “remember” the interpretations combined with human feedback for later analyses.

  1. No AI = no cost savings realized

Gualtieri emphasized the critical impact of AI on the ROI of retailers’ analytics efforts. He cited a past Forrester survey in which large retailers were asked if they could quantify the positive benefits of their analytics efforts. Shockingly, only 12% of them said they could do this. This was a testament to a major limitation of traditional data-analysis and business-intelligence solutions — they don’t generate returns because their insights aren’t actionable.

Gualtieri used the example of Tableau, a common data-visualization tool: “You have to pay someone a big salary to sit in front of it and make pretty pictures about stuff that happened…months ago,” he said. “At Forrester we call these ‘perishable insights.’ If you can’t act on [it] in the time that it happens, what’s the point?”

I couldn’t agree more with his points. We hear this sort of thing all the time here at Profitect. New prospects often ask us, “So if I give you our data, can you find things about my business that I’m not aware of?” The short answer is absolutely. But prescriptive analytics can also tell you what to do about those problems — and that’s where the real monetary value appears. Finding unknown things is great; doing something about them is even better.

With its striking ability to quickly turn raw data into actionable insights and opportunities, AI is unquestionably the future of data analytics. To learn more about how AI can help you see next-level cost savings at your company, visit Profitect at www.profitect.com or contact us at info@profitect.com.