Zones Execs: ‘Most Customers Think AI Is A Magic Wand’

“[Customers] click ‘Subscribe’ on Copilot and now they’re ‘doing AI.’ But they don’t understand what Copilot is actually doing. … So what we do is help them map it out. We show them what’s happening under the hood, and we start small,” Brad Davenport, senior vice president of security and architecture at Zones, tells CRN.

In the race to implement AI across all industries, it’s not about using the latest and greatest tools, but providing value to customers that are starting from scratch.

The value comes from clarity, collaboration and meeting clients where they already feel confident. That’s not just good AI strategy; it’s good business, according to Tony Berg, senior vice president of solutions and architecture at Auburn, Wash.-based solution provider Zones.

“I think we’ll always have some level of curiosity, but it’ll evolve. It’s like anything else. At first, it’s ‘What is this?’ Then it becomes ‘How do I use it?’ And eventually, it’s ‘How do I optimize it?’” Berg told CRN. “That’s the curve we’re seeing now. People are getting familiar at home. Now they’re trying to make it work at work. And the companies that succeed will be the ones that don’t just chase buzzwords, but actually simplify their problems and start solving them, one use case at a time.”

AI isn’t any different than what came before it, it’s just the next big leap, according to Brad Davenport, senior vice president of security and architecture at Zones, No. 38 on CRN’s 2025 Solution Provider 500 list.

Davenport (pictured above) shared insight about the realities facing IT leaders today, from talent shortages to growing system complexity and why cutting through the noise is more critical than ever.

“In 10 years, AI will be just like cloud computing is now: table stakes. There’ll be a Microsoft, a Cisco, a Salesforce of AI and a bunch of niche players,” Davenport said. “The market will consolidate, and customers will be locked in. But even then, it’s still going to be complicated. There’s still going to be problems. There’s still going to be “I don’t know how this thing works” moments. And that’s where we come in.”

CRN spoke further with Berg (pictured just above) and Davenport about the complexities around AI and automation and how they’re bringing value to customers.

When talking to customers, what is the biggest challenge they have today?

Berg: One of the biggest challenges right now is complexity and just how intricate and layered their environments have become. Everyone’s hybrid now. They’ve got on-prem, they’ve got cloud, they’ve got SaaS apps and [are] securing all that without making it harder for end users or customers. That’s a big hurdle. It’s not just about tools; it’s about making all these moving parts work together without breaking the experience for users.

Talent is another one, hands down. There’s just not enough of the right people. Security is so broad now, someone might know identity and access management but have no clue how to properly secure a cloud environment. The talent pool is either too shallow or the people they have don’t have the breadth of knowledge they need. And finding partners to supplement that skill gap? That’s a massive challenge too.

Davenport: It’s almost always about acceleration—how do we get this done faster? And that’s regardless of what part of the tech stack we’re talking about, whether it’s infrastructure, cloud, AI or security. Customers have made their decisions, secured their budgets and often waited 12 to 18 months for gear to land or software to get provisioned. By the time we’re implementing, the original business drivers might have shifted. So the first question out of their mouth is usually, ‘How fast can you make this real?’

But here’s the thing, implementation is just the beginning. The full ROI of a solution doesn’t arrive the day it’s installed. It’s iterative. It needs provisioning, adoption, user education, tuning. And then it still needs continuous engagement. So whether you’re a product vendor or a service partner, customers want to hear how you're going to handle the full life cycle from procurement through implementation to long-term value realization.

So how is Zones untangling that for clients?

Davenport: It really is a giant machine. You’ve got micro systems, interdependent processes, outsourced elements and multiple departments handing off responsibilities like a baton in a relay race. And when we show up and say, ‘Let’s automate this,’ it sounds simple on a whiteboard. But the moment we ask the customer to walk us through the process, step by step, they hit blind spots. They literally don’t know how a certain report gets created or who touches it before it goes live. ‘We put in a coin and a widget comes out’ is often the level of understanding they start with. So our role becomes less about selling a tool and more about initiating a self-discovery process. We whiteboard with them. We identify where the real pain points are. We say, ‘Hey, instead of trying to automate your whole data center, what if we focused on just this one workstream that’s causing the most drag?’ And once you digitize that, now you can monitor it, collect logs and generate the telemetry that opens up visibility across the business.

So are customers leaning on automation and AI to fill those skill gaps? Is that part of the conversation you’re having with them?

Berg: Yes, absolutely. But there’s a catch. Automation can be a double-edged sword. If the people running it don’t understand what the automation is doing or don’t trust it, they hesitate. There’s a fear factor there. Like, ‘I don’t want to hit this button if I don’t know exactly what it’s going to do.’ So part of our job is not just implementing automation but making sure they understand it.

Are those conversations still in the early, exploratory phase or are some customers already trying to scale it?

Berg: Most are still ‘AI-curious’—that’s actually a term we’ve started using. They're dipping their toes in things like Microsoft Copilot for meeting notes or action items. It’s a good start. But then we help them dig deeper and ask questions like, ‘What data do you have?’ Because if your data isn’t structured right, your AI’s not going to be useful. That’s where a lot of people get tripped up. They have data all over the place, siloed off behind firewalls or in legacy systems that aren’t AI-ready.

And how are those conversations evolving?

Davenport: The first challenge with AI is maturity. We ask, ‘Where are you with data governance? Automation? Do you even know where your PII [personally identifiable information] lives?’ Because most customers think AI is a magic wand. They click ‘Subscribe’ on Copilot and now they’re ‘doing AI.’ But they don’t understand what Copilot is actually doing. It’s crawling their SharePoint, their email, their SQL databases. When compliance officers realize that, the conversation gets real fast.

So what we do is help them map it out. We show them what’s happening under the hood, and we start small. Instead of building a full AI model, maybe we automate employee on-boarding. That’s a controlled scope. Then we move to document classification, then service desk response optimization. It’s step-by-step maturity.

Berg: We work with them on two fronts. First, there’s a bit of data reality-checking: what they think is clean, structured data often isn’t. So we help validate that. Second, we lean into vendors and OEM partners. A lot of them have already done the integration work, which lets our customers move faster. But the key advice we give: Don’t force your tools to fit your process, change your processes to fit the tools. You don’t need 18 approvals to do something just because that’s how it’s always been done. If the tool only needs three, ask why you’re making it more complicated.

That mindset shift seems huge. Are you finding customers are open to it? Or do legacy mindsets get in the way?

Berg: It’s definitely a shift. But one thing that’s helping is how fast consumer tech has moved. People have AI in their phones now. They’re using Siri or ChatGPT to plan vacations, write emails, even do schoolwork. So now, when they see similar tools at work, they’re not as intimidated. They’re more willing to try. That comfort at home accelerates adoption in the enterprise.

Speaking of ChatGPT, are you seeing it show up in unexpected ways?

Berg: Oh, definitely. I had a client tell me their HR screeners were noticing all the job candidates answering questions the same way. They started suspecting candidates were using ChatGPT in real time, like with their phones in ‘listen’ mode during interviews. So that client actually started using ChatGPT themselves to ask the same question and see if the answers matched. That’s cybersecurity in a weird way. You’re looking for anomalies, for digital countermeasures.

So once customers move past curiosity, how do you help them get to real impactful AI use cases?

Berg: We always say narrow the scope. I had one customer come in with 78 use cases. That’s not practical. We helped them pick the top four—ones that would show ROI without requiring a team of 25 data scientists. Start small, get some wins and then expand. And those wins have to be simple: automate meeting follow-ups, summarize action items, improve search in support tickets. Things that are tangible.

So where do your vendors come in? What do you need more of from them to make this all work better?

Berg: Innovation and integration. AI isn’t just a software problem, it’s a power, cooling and infrastructure problem too. So as vendors innovate to make AI more efficient, smaller, faster, cooler, that’s going to unlock new possibilities. But also, we need them to stay open and collaborative. That kind of openness is what’s speeding everything up.

Davenport: We’ve been pushing hard on our partners—Cisco, Microsoft and others—to de-obfuscate AI. If your tech has an AI layer on it now, and let’s face it, everything does, then give us, the partners, the tools to explain it to customers. There should be something like a panic button in AI-driven systems, a red light where if something goes wrong we can immediately see why it made that decision. Just like in traditional systems, there needs to be a manual, a triage flow. You have to be able to debug it. Otherwise, people won’t trust it.

Do you see customers moving from ‘AI curious’ into full AI adoption sooner rather than later, or will some stay stuck in that curious phase?

Berg: I think we’ll always have some level of curiosity, but it’ll evolve. It’s like anything else. At first, it’s ‘What is this?’ Then it becomes ‘How do I use it?’ And eventually, it’s ‘How do I optimize it?’ That’s the curve we’re seeing now. People are getting familiar at home. Now they’re trying to make it work at work. And the companies that succeed will be the ones that don’t just chase buzzwords, but actually simplify their problems and start solving them, one use case at a time.

Davenport: When we go in we ask, ‘What’s the most resilient process you’ve got? The one that's like clockwork, with clear ownership and lots of historical data?’ That’s where we start with AI. Why? Because the business already trusts that process. They know the expected outcomes. If we can make that even better using AI, it’s a confidence builder. Now you’re not gambling with an unknown; you’re enhancing something solid. That’s how you get AI adoption to stick.

I have heard about the economics of automation, especially for MSPs. There seems to be a tension there. Can you unpack that?

Davenport: Oh yes, big time. So, say you’re an MSP and it traditionally takes 100 man-hours to manage a certain task. Now, thanks to automation and AI, you can do it in 70. Sounds great, right? But here’s the catch, how do you go back to your customer and still charge the same price? From their perspective, if it takes less time, it should cost less. But for us, we had to invest up front and analyze the process, build the automation and test it. That’s not free.

And here’s where it gets weird: Some MSPs don’t even tell the customer they automated the process. Because the moment it looks too easy, the customer starts questioning the value. So they have to hide their efficiency just to preserve pricing integrity. It’s like this quiet conundrum we’re all stuck in.

So the fear is that showing how efficient you are actually hurts you financially?

Davenport: Exactly. It’s like being punished for doing your job too well. I mean, look at cable companies—your bill doesn’t go down when it gets easier for them to deliver content. You still pay the same. But in our space, if we’re not careful with how we frame the value, customers start thinking, ‘Why am I paying this if it’s all automated?’ So we’ve had to learn to pivot the conversation. It’s not about time anymore, it’s about business value.

What does that look like in practice? How do you turn an efficiency gain into a value proposition?

Davenport: You decouple the conversation from the tech. Let’s say you’re in retail and your defect return rate is 20 percent. Instead of saying, ‘We’ll use AI to automate this process,’ we say, “What if we could reduce your returns by 5 percent? That gets products back into the refurb market faster. That boosts your margin.’ That’s a value prop. It’s not about how many hours we saved; it’s about what outcome we helped create. That’s how you protect your pricing and, frankly, your worth as a partner.

Do customers get that? Or is there a gap between SMBs and larger enterprise customers in how they perceive that value?

Davenport: Totally. The smaller the customer, the more they're just focused on cost. SMBs, especially the one- or two-man shops, they just want to spend less. But if you’re in enterprise or government, they care about value and risk. They understand ROI. So it is a dance. You’ve got to know your audience. But ultimately, whether it’s a small business or a Fortune 500 company, people want to feel confident they’re making a smart investment. And AI is still so new to most that they lean heavily on us to make that confidence happen.