The Three Legs Of AI: A Framework For Building Successful AI Systems

InnovationThe Three Legs Of AI: A Framework For Building Successful AI SystemsByMatt Shea,Forbes Councils Member.for Forbes Technology CouncilCOUNCIL POSTExpertise from Forbes Councils members, operated under license. Opinions expressed are those of the author. | Membership (fee-based)Jun 10, 2026, 09:00am EDTMatt Shea is Chief Strategy Officer at MixMode. gettyDeterministic. Statistical. Contextual.There are three "types" of AI needed for effective AI systems, the proverbial three legs to the stool.LLM/GPT technology (Claude, ChatGPT, etc.) is being considered for a wide range of applications. Many breakthroughs are being enabled, but the trap of viewing them as the solution for every problem should be challenged.This is a classic cognitive bias called the "law of the instrument," also known as the "golden hammer." When you have a golden hammer, it's a mistake to view all problems as nails. Basically, you shouldn't smash a screw into the wall with a hammer.The Defense Advanced Research Projects Agency (DARPA) gave us the template for understanding AI systems when they published the Three Waves of AI in 2017. The presentation covered 60-plus years of AI advancement and presciently predicted the arrival dates of new technology.DARPA spelled out that the first wave of AI was embodied in expert systems. Long since stripped of the AI moniker, these are deterministic in nature, where the same inputs always produce the same outputs, with logic handcrafted at significant complexity. The classic example is TurboTax, where tax attorneys and programmers turned the IRS tax code into an expert system that calculates taxes due, updated each year. In cybersecurity, an example is the rules put in place to find known cyberattacks from existing signatures.In short, deterministic logic is "if this then that" computation. Fixed logic that can't learn or adapt on its own.There are many problems where writing the logic by hand is to...