Why Tech Vendors Must Define Their AI Category Now …Before Competitors Do It For Them
By Peggy Tierney Galvin, co-founder, chief strategy officer
Before I could write this blog, I had to brush branches, leaves and fiber optic cables out of my hair and sweep the pile of NASDAQ ticker tape off my desk. All of us tech comms pros are still a little wild-eyed from the torrential weekend of AI news earlier this month.
To wit: AI leaders declare grave danger to humanity, beg government to rein them in! (One imagines Sam and Dario as Frog and Toad, guzzling cookies while declaring the need for willpower to avoid guzzling cookies.) U.S. government responds saying, AI problem? What AI problem! European governments declare AI regulation necessary! Chronically online people stammer, “what do you mean, ‘grave danger to humanity?’” Contrarians point out economic self-interest from frontier models in arresting the rate of competition prior to IPOs! Economists note that the economy would crater if AI development slowed down!
Fig. 1: Scientific rendering of the discourse
As a layperson fortunate to know many people with insight on this topic, I asked John Hendley of Coalfire for his take. On a Monday morning huddle, John noted that the eyewatering heights of frontier model performance are a textbook example of diminishing gains. More pointedly, an exorbitant amount of investment capital is being spent to eke out only slight improvements on existing models, which are more powerful than even many edge cases would need. For many – frankly, most – enterprise users, the “Mazda” open-source AI model is perfectly fine. Enterprises are only too happy to use these affordable, open-source AI models, while big companies with specialty needs are already experimenting with banks of Nvidia servers in-house, tinkering with their own open-weight models, bypassing the Sams, Darios and Alexes completely.
So, what does this mean for tech vendors looking to sell their solutions to enterprises like these? Surprisingly, there’s a glimmer of optimism. In a market where AI-native solutions are galloping to cordon off market share from legacy solutions, inches away from the legacy solutions racing to AI-wash their offerings in the latest lingo, the winner will be the company that can define their customer’s business pain clearly. It will be the company that can accurately and confidently describe how their use of AI delivers ROI and that tells their message consistently in all channels where their audiences are looking for help. (End of the world notwithstanding.)
