Despite security vulnerabilities and cost concerns, businesses are accelerating AI adoption. The Global Technology Industry Association (GTIA) warned at its annual ChannelCon event that IT service providers risk being left behind as customers rush into AI. Partners are feeling the pressure.

"I'm receiving many customer projects that have become extremely urgent because they've discovered new AI tools they want to implement," said Jeremy Grissom, professional services manager at STR Technologies. "It's full speed ahead."

Cloud computing and internet technology are useful analogies for how AI adoption is reshaping the managed services landscape. It's just that AI is moving faster.

"We're going to have to start addressing these issues more quickly," said Carolyn April, GTIA's vice president of research and market intelligence, at a press conference last week. "Managed services adoption progressed very slowly. I don't think we can move at that pace with AI... That's the reality."

The industry association launched several initiatives last week aimed at developing industry standards to protect AI and managed IT services security.

GTIA also announced the formation of the "Managed Intelligence Alliance," which will develop standards, certifications, and specifications for partners managing AI services. Pax8 initially promoted the concept of "managed intelligence" but had to push it to a third party.

"Although I think my organization is great, we're not neutral," said Rob Rae, vice president of community and ecosystem at the cloud commerce platform, at the same press conference. "Some MSPs won't use us. This idea is bigger than us, bigger than Pax8."

GTIA is already poised to become the industry association partner for the "Responsible IT Services Alliance," recently launched by Texas A&M University's Global Cyber Research Institute. The organization includes Pax8, New Charter Technologies, and The20, aiming to establish IT industry norms before government intervention.

"Regulation is coming, either imposed on us or we build it ourselves," said GTIA CEO Dan Wensley. "We have the resources to do this. We have industry expert members. We have to do it."

Jennifer Gilligan, president of IntegraMSP, believes government officials won't necessarily mandate AI compliance. She suggests cybersecurity insurance companies will play a leading role.

"Insurance companies are the ones bearing the front-end pressure of AI-related claims," said Gilligan. "They've already started injecting these questions into renewals. What's your stance on AI? Are you using AI? Do you have an acceptable use policy? This is already beginning."

Two paths for AI

Last year, when IntegraMSP faced the choice of building AI or protecting AI, the company chose the latter.

Gilligan said the company lacked a developer team to build AI tools for clients. Instead, IntegraMSP leveraged its strengths in cybersecurity.

"It felt like building AI would become a race to the bottom because there will always be someone faster, cheaper, and at larger scale than me, whereas I've already built a solid foundation in security protocols," said Gilligan.

Other partner companies have taken a different approach with customers who expect MSPs to build and manage AI tools.

"If you encounter a large customer like in California, they might ask you to train a large internal model for them, but I'm not sure that can be done by an MSP," said Jack Skinner, president of Oversee My IT. "It depends on the customer, and it depends on the size of your organization."

IntegraMSP first conducts an AI readiness assessment to evaluate the state of customer data and determine which LLMs their employees prefer and use. Then they develop an acceptable use policy that customers can submit to their insurers and use for employee training.

Gilligan said the goal is to go beyond the initial project and continue accompanying customers to help them sustain AI adoption.

"The real value lies in the consulting services you provide on top of the tools," said Gilligan. "Anyone can sell tools, but how to use them and drive discussions from a whole-business perspective rather than just a technical one—that's the key."

One of the first steps MSPs take in AI governance is preventing customers from entering company data into public chatbots. Reports that shared Claude chat logs appeared in Google search results have heightened concerns. Craig Fulton, M&A advisor at Evergreen Services Group, warned MSPs that commercial AI models are not a safe place to share M&A roadmaps.

"I call it a five-year-old who hears everything you say as an adult," he said. "Then they repeat certain things. You think, 'Where did you hear that?' 'Daddy, I heard you on the phone.'"