How Do We Win in an AI World? (Part 1)

 

Is the bloom off the AI rose?  AI optimism seems to be collapsing under a black wave of AI gloom.  To be sure, the effects of AI are everywhere, both good and bad, though the bad tends to grab the headlines. My practice is focused on improvement & innovation– both Digital & analogue.  What am I seeing & what does it mean?

Companies I work with are facing the following conundrum: How do we harness the power of AI while avoiding the jagged rocks dominating the headlines? The pushback against AI is real & growing.  People are rightly worried about their jobs, as industry after industry comes under attack. (And are humans not increasingly repelled, especially by AI in disguise, pretending to be human?)

Let’s imagine we are a large multinational customer-facing company. We have an honorable history of service to our communities, team members and shareholders. We’re under intense competitive pressure & are facing daunting Innovation, Lead Time & Cost challenges. Like so many of our peers, we’ve invested heavily in AI and know the Board expects a rapid ROI. And yet, our team members, though loyal & capable, are understandably hesitant to apply AI tools to improve productivity, reduce & otherwise improve our business. What to do?

AI’s challenge is not unprecedented. In fact, it reminds me of the challenge presented by Robotic Process Automation (RPA) a decade or so ago when it began to be applied at scale in industries like Financial Services. RPA is deterministic AI (If X, then Y) and has been used to automate data entry, invoice processing, and other high-volume, routine tasks. Agentic AI is non-deterministic (If X, then Y or maybe Z or maybe W) and excels at complex, cognitive tasks like customer support, strategic planning, or managing end-to-end IT operations. Unlike RPA, Agentic AI makes decisions based on your needs & the situation on the chessboard. Moreover, and this is perhaps what scares people most, Agentic AI learns through repetition. This is why chess programs like Leela and Alpha Zero continually get stronger, so much so, that human World Champions struggle to secure even a draw against them.

Many of the lessons we learned with RPA apply to Agentic AI. So, how do we win in our brave new world? Let’s begin with the most basic point: Agentic AI, like RPA, is a tool. As such, it can be an good servant but is a dreadful master. Let’s therefore, in every use case, define the problem we are trying to solve, and only then consider what tools might suit. As we’ll learn in coming articles, Agentic AI is rarely the first too you reach for.  Stay tuned.

Best wishes,

Pascal Dennis         E: pascal.dennis@leansystems.org