Hey AI, time to earn your keep!

Not too long ago, there was a general sentiment of FOMO in organizations with regards to AI technology. The value of knowledge embedded in data to help drive better, more informed decisions had been well-established; but many organizations were not equipped with appropriate talent or tools and perhaps did not have sufficient data acquisition and retention practices in place to make this a reality. It was common to see a bit of panic set in with the fear that others might be gaining a real competitive advantage using AI, or that as an organization they were missing an opportunity to modernize how they do business or how they relate to their customers.

So, data acquisition and retention became a major priority, along with the mad scramble for data science talent - a race to acquire people with the knowledge and skills to put AI technology into practice. In retrospect, with such urgency this was often done quite blindly, with little regard for identifying the specific problems to address that will be of real value.

Colloquially you might say it was a lot of “hey smart guys, here’s a bunch of data – find us some golden nuggets.” Organizations often had little to show for their AI investments because of the hyper-focus on model building and model performance, on exciting innovation over practical application. Even when the business objectives were well-defined and AI models showed great promise, operationalizing them was often just deemed too disruptive. Thus, the payoff from AI in terms of real outcome was still yet to be fully realized.

The time has come! We, at SAS, have been afforded a ringside view of how organizations have experimented with AI and we are now seeing the maturity seeping in with many organizations keenly looking to take AI from the lab into the real world. In general, organizations have moved beyond the phase of just getting their feet wet with AI. Investments in AI – in terms of both technology and people – are ready to pay off.

Largely, it’s a maturation of their “digital transformation,” but as with most big changes these days there might be a bit of a Covid-19 factor here, a reset of sorts. With disruption as the status quo and historical patterns no longer necessarily reliable, what better time to re-evaluate, re-tool, and re-implement, making AI an integral component of their business operations. Now is the time to replace or augment traditional processes bound by human limitations with AI-driven processes that can exploit latent intricate relationships in high-dimensional data to offer deeper insights and optimal recommendations. The rapid maturity of AI orchestration initiatives is evidence of the willingness and desire of organizations to abandon the status quo and embrace AI as an integral component of their operating model.

I know, many organizations have had well-established AI-based processes in place for years. And, we all interact with services backed by AI every time we swipe our credit card (fraud detection), use a mapping app to get driving directions (optimal routing), or perhaps unlock our phones with face ID (facial recognition), to name a few. Here, I’m referring to more of an across-the-board integration of AI technology in every aspect of how an organization operates – not only for unique breakthrough projects, but also to realize value by applying AI techniques to established projects and processes to achieve best-in-class results.

For an AI product or service to be successful, it will incorporate elements that will help make an outcome better, or a process faster or cheaper. Consider a customer service interaction that is tuned to achieve optimal customer satisfaction through analytically driven decisions for processing product returns or exchanges, with individualized experience and offers for the customer and optimized logistics and inventory management for the retailer. Or, think even bigger!

We’ve all heard at least one story of AI gone wrong, and the ensuing justified backlash. But think of the dramatic impact that organizations can collectively have on society as a whole by using AI to identify and alter behavioral patterns that inadvertently instill and perpetuate bias into existing systems and processes. You must identify and emphasize a problem before you can fix it…and AI provides that opportunity.

The comfort and maturity level that organizations have with AI have reached a point where they can expect to see their investments in AI pay off through improved services and offerings that deliver real value to customers. Certainly, they won’t all be success stories, and failures will continue to make headlines. But mainstream adoption will overtake research projects, with consumers, and perhaps society in general, as the beneficiaries. The value of AI will be determined not by how well it models the real world, but by how it helps improve it.

-- Brett Wujek, Principal Product Manager for Analytics, SAS.