We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – August 3. Join AI and data leaders for insightful talks and exciting networking opportunities. Learn more about Transform 2022
Stories of how AI will benefit the enterprise are a dime-a-dozen these days. Applications in sales, marketing, payroll and a host of other areas are legion. But as of yet, there is precious little talk about how, exactly, organizations are faring with their AI projects. Are they really delivering on these promises, and might there be some concrete examples of AI at work that can be emulated elsewhere?
Judging by Gartner’s Hype Cycle, most organizations are close to wrapping up the developmental and experimentation phases of their initial AI programs and are now looking to operationalize them within the business model. This is a crucial step for the technology because it represents the leap from expectation to reality. Without tangible results in the real world, such as heightened productivity, lower costs or some other positive outcome, AI could get pushed back to the lab for further refinement or possibly suffer a slow death altogether.
Positive outlook for AI
According to MIT Sloan Management Review, however, 2022 is shaping up to be the year that AI finally starts to produce solid returns on the investments of the past few years. In 2019, for example, only three out of 10 surveyed companies reported even minimal value from their AI endeavors, with failures largely attributed to the difficulty in pushing the technology into production environments. This year, more than 90% are reporting solid returns on their AI investments and are planning to expand their strategies going forward.
Surveys are all well and good, but where are the success stories that highlight exactly who is benefiting from AI, and how? Investment firm Vanguard is one such example. Its retirement plan division, Vanguard Institutional, needed a way to deliver information, service offerings and other content to customers not just in a general fashion but on a personalized basis, and at scale. Using a natural language platform developed by Persado, the company can now target individual customers with exact phrasing, formatting and relevance, leading to a 15% increase in conversion rates.
Marketing, in fact, seems to be the initial hotbed of activity for AI in production environments. Discite Analytics and AI recently posted five examples of how companies are using the technology to distinguish themselves in an increasingly crowded and noisy business environment. Lays, for one, recently utilized deep fake technology to allow users to customize video messages of Argentinian soccer player Lionel Messi to share among friends. Mattress firm Tomorrow Sleep put AI to work on their content marketing programs to identify ways to improve organic traffic – and saw a jump from 4,000 visits per month to more than 400,000 per month within a year.
Operational balance your McDonald’s order
AI’s capability to optimize the supply chain is also starting to come into focus, not just in terms of product sourcing and distribution but in customer-facing operations as well. McDonald’s recently acquired an Israeli company called Dynamic Yield that provides personalization software that can do everything from recommend food and beverage offerings at the ordering kiosk to anticipate traffic volumes and preferences based on multiple criteria, such as weather and public events. At the same time, however, it can read inventory levels to promote items that are plentiful and discourage items that are scarce, bringing supply and demand into greater alignment in a highly dynamic fashion.
These are just some of the ways in which AI is being put to practical use. Undoubtedly, the technology still has a long way to go before it enters the economic mainstream, and there will surely be far more examples of AI failures than success for the time being.
But the distinguishing factor between AI and previous forms of digital technology is its capability to adapt to changing circumstances. This means that when it fails, or simply doesn’t fulfill expectations, it can easily be retrained to produce a more optimal outcome — no more going back to the drawing board for a complete code rewrite that may or may not address a problem that might not even be relevant anymore.
In the digital world, AI makes possible the old adage “If at first you don’t succeed, try, try again.” And even when that threshold of success is finally met, AI can be continuously refined to drive that success to higher and higher levels.