AI & AUTOMATION How MSPs move AI from hype to business value By Michael Gray The Four Pillars of Managed AI To date, AI projects have been a big, expensive experiment. Innovators and early adopters jumped on the hype and have shelled out billions collectively to test and stretch AI to see what’s possible. All things considered, success has been limited. Nearly 90 percent of AI projects never make it into production, according to Garnet estimates, and 92 percent of enterprises reported at least one failed proof-of-concept in 2024. Despite how bleak that sounds, these experiments haven’t been entirely fruitless. In fact, they’ve given us incredible insights into this technology in record time. But we’re moving further into the adoption lifecycle, where midmarket pragmatists and the greater mainstream market are starting to pick up the mandate. This group is notoriously more hesitant than the early adopters that came before them. We’re already seeing it play out with midmarket companies that are now far less whimsical in their AI approach: we’re seeing less emphasis on experimentation at any cost, a bit more return consciousness and decision-makers more influenced by data than rhetoric. These early experiments made it clear that the problem isn’t the technology; it’s the lack of a plan. And that hasn’t completely been a bad thing. For a lot of these experiments, failure was the point. Fast failure guides 12 CHANNELVISION | FALL 2025
RkJQdWJsaXNoZXIy NTg4Njc=