MSPs Are Driving GenAI Adoption But Security, Data Hurdles Persist For Customers: Panel

While MSPs are showing the way forward on generative AI, many customers end up needing to step back and ensure that data readiness and security risks are addressed upfront, according to experts who spoke at XChange August 2024.

While MSPs are helping to show the way forward on generative AI adoption, many customers at some point end up needing to step back and ensure that data readiness issues and cybersecurity risks are addressed upfront, a panel of IT channel executives and experts said Monday.

The surging interest in GenAI tools such as Microsoft Copilot has created major new opportunities for MSPs to show their value as trusted advisors, while at the same time creating challenges in situations where customers might be looking to move too fast, the panelists said during the session, which took place at XChange August 2024. The conference, held this week in San Antonio, is hosted by CRN parent The Channel Company.

For instance, if you roll out Copilot, “you need your security team” to be heavily involved, said Julia Deien, partner and technology practice leader at St. Louis-based Anders CPAs and Advisors. “Because [security] is a big question mark for all of our clients.”

During the panel, moderated by The Channel Company CEO Matt Yorke, panelists discussed findings from a recent AI study from IPED, the channel consulting arm of the company. These findings from the survey of solution providers highlighted the fact that cybersecurity is currently seen as the biggest inhibitor for AI adoption, followed by concerns about the technology’s accuracy.

When it comes to security concerns over GenAI usage, “no not only is it [about] hacking — it’s data privacy,” said Mark Williams, senior consultant at IPED.

As a corollary to this, however, the survey found that security and risk management is considered the largest solution offering opportunity for solution providers over the next 12 months. That’s followed by software and analytics services, with platform and data management services coming in as the third-largest opportunity in the next 12 months, according to the IPED study.

Without a doubt, the readiness of an organization’s data for usage with AI tools like Copilot continues to be a massive issue, said Don Monistere, president and CEO of McLean, Va.-based General Infomatics.

While many customer conversations have started out focused on AI, “it's really interesting how the conversations are changing [to be] about data and data cleansing,” Monistere said.

While General Infomatics has been aggressive about seeking to enable usage of new GenAI-powered tools in its customer base, the conversations with customers are frequently focused at this stage on issues such as whether the permissions for accessing SharePoint data through Copilot are correct, he said.

“Those are the conversations that we're typically having,” Monistere said. “So it's more about being AI-ready, as opposed to truly embracing the AI tool sets that we've actually provided.”

Deien noted that once Anders’ customers become aware of how much work has to be done to clean up their data, it can sometimes lead to a reluctance to move forward.

“They see the value, but they don't want to spend the money, because they know their data is dirty,” she said. “If you want to clean up the data in their production [systems], that's what it's going to take.”

While Copilot is “great,” the time quickly arrives when it’s going to touch the data within live production systems — and for many organizations, that is going to require hiring additional specialists, Deien said. The big question then becomes, “how do you start investing in that area with the right people?”

There’s no question that for many organizations that are eager to get started with GenAI, “not enough time is given to securing your data first,” said Dennis Cockrell, founder and CEO of EIT Networks and Cybersecurity in Greenville, S.C.

“We've refused to actually deploy [AI] for our clients until we have that data analytics conversation — looking at data security, classification,” Cockrell said. “I think it's kind of irresponsible for folks in our space to partner with companies that just want to go turn it on, without understanding the risk. It's our job to educate them and to help them understand that risk.”

Likewise, many MSPs are still at the stage of needing to “get our bearings and understand the ramifications — [such as] what does a breach look like in AI? And what does that really mean to our clients?” said Atul Bhagat, president and CEO of BASE Solutions in Vienna, Va.

Ultimately, “we have to be very careful in what we're feeding it — what information are we giving it — and where that information is going, where it’s being stored,” he said.

At the same time, MSPs such as General Infomatics have been leading by example when it comes to utilizing GenAI for increasing productivity and efficiency. The company has developed its own AI chatbot, dubbed “Megan,” which the company has been using to answer questions from employees — and even to field some external inquiries, Monistere said.

“When new hires come along, we tell them all the time, 'Go talk to Megan.' And they don't actually know that she's not real,” he said. “She knows everything about about [the company]. So a lot of times [when] they ask Megan, she typically has the answer.”