By Sarah Nicastro, Founder and Editor in Chief, Future of Field Service
While I’ve been on the receiving end of plenty, I’m not personally one to throw around a sports analogy too often.
However, when I was talking with Dr. Aymen Gatri for a recent episode of UNSCRIPTED, the conversation about AI risk brought one to mind that I couldn’t shake (or pass up sharing).
My eldest son, Evan, plays basketball. He absolutely loves the game, and over the past few years I’ve found myself repeatedly reminding him of something that isn’t always especially exciting to hear when you’re young and focused on scoring:
Offense gets the glory, but you can’t win games without strong defense.
You shoot a basket, it goes in, and everyone celebrates. It’s visible. It’s exciting. It’s easy to point to the scoreboard and say, There’s the impact.
Defense is different.
Good defense can be less noticeable. It’s about positioning, anticipation, awareness, and discipline. It’s often about preventing the other team from capitalizing on an opening, but strong defense doesn’t only protect your team – it creates opportunity. The challenge – especially for young kids – is that sometimes, the value of a defensive play is only obvious because of what didn’t happen.
(As a quick sidebar, I am so passionate about this point that I’ve been known to yell from the bleachers (sometimes very passionately), “Rebounds win games!” We took the boys to see a Cleveland Cavaliers playoff game this past season, and there as an area where you could make your own signs. My sons quickly said, “Mommy – you should put ‘Rebounds win games!’” It was a nice reminder that they do listen.)
So, it dawned on me during my conversation with Aymen, and I’ve thought more about it since, as a useful way to think about where we are with AI in service.
Right now, AI offense is getting a lot of attention. And for good reason; the potential is enormous.
Aymen wanted to be clear in our discussion that, while he wants to draw more attention to the risk that’s inherent in AI that he feels many are overlooking, he isn’t at all trying to diminish the excitement that the potential of AI brings. He feels it, too.
But if service organizations want to capitalize on that potential sustainably, they need to give just as much thought to their AI defense as they are their offense.
6 Aspects of a Strong AI Defense
One hypothesis I have is that leaders know they need a stronger AI defense, but they haven’t yet determined exactly what that looks like. In the rush to keep pace, offense is getting more focus not only because it gets the glory, but because it’s the clear starting point.
Aymen weighed in with his thoughts on what defensive concepts service organizations should consider augmenting their offensive approach with:
1. Start with strategy, not the technology.
One of the most important points Aymen made during our conversation was that the initial risk is implementing AI simply because AI is happening (without clear objectives).
That sounds obvious, but I think it's increasingly difficult to avoid the pressure.
There is a new capability, a new model, a new application seemingly every day. Organizations know they need to keep pace, and that urgency can easily turn into a mentality of racing rather than reasoning.
But the starting point should be intentional. Put very simply: What problem(s) are we trying to solve?
Aymen suggests starting the conversation with a fundamental question: Why does service matter, and why does AI matter for service?
The underlying objectives of service haven't suddenly changed because AI exists. Organizations are still trying to get things fixed, reduce downtime, increase uptime, improve customer outcomes, and address the very real challenge of getting knowledgeable people where they are needed.
AI can help with those challenges. But it helps most when the technology serves the strategy — and often goes awry when technology becomes the “strategy.”
A strong AI defense therefore starts before implementation, with clarity around what the organization is trying to accomplish and where AI genuinely fits.
2. Establish clear governance and accountability.
Aymen was sure to underscore that the risk profile changes significantly when we move from AI that assists people to AI agents that can take action. As we move toward systems that can ask questions, reason, make decisions, and potentially conduct work on our behalf, he suggests organizations seek answers to some fundamental questions:
- Who governs what?
- Who owns what?
- Who is accountable for what?
- And can an AI agent itself be held accountable?
These aren't necessarily questions with simple answers today. In fact, part of the challenge is that we're moving into territory where many organizations haven't yet established the appropriate frameworks. That's precisely why governance can't be an afterthought.
Aymen compared this evolution to cybersecurity. As cybersecurity became increasingly important, responsibility expanded beyond traditional IT security into areas including OT and IoT, ultimately creating a need for broader organizational governance.
He sees a similar requirement emerging with AI. Whether that means a dedicated AI leadership role, an existing executive taking on the responsibility, or some other structure will vary by organization.
There isn't one model that fits everyone, but what matters is that there is a clear home for AI governance, accountability and risk.
3. Build risk and compliance into the AI strategy.
A strong AI defense isn't about identifying every possible risk before moving forward. That’s not possible. The technology is developing too quickly, and there are questions we don't even know to ask (and more to come). But that doesn't mean risk should be left out of the equation.
Aymen's recommendation is to incorporate AI risk and compliance directly into accountability, risk, and strategy planning. He pointed to the evolution of cybersecurity and the established management systems organizations have developed around quality, information security and other areas as useful parallels.
The idea is essentially to create a management approach around AI that allows an organization to pursue the opportunity while deliberately managing the risk. That distinction is important. Risk awareness isn't an argument against AI adoption; it's part of responsible AI adoption.
4. Know your data — and protect it.
For service organizations, this becomes particularly important because AI may operate across an ecosystem involving the service organization, its customers, suppliers, and technology providers.
Aymen raised a deceptively simple question: Who owns the data? Customer data. Corporate data. Supplier data.
Also: Who is processing it? Where is it going? What infrastructure sits behind the AI agent? What happens to that data throughout the lifecycle?
Those boundaries aren't always clear. And when an AI agent is acting on behalf of an organization, the implications of unclear data ownership and processing can extend beyond a technical concern. There are privacy implications. Cybersecurity implications. Intellectual property implications. And, ultimately, business and reputational implications.
A strong AI defense therefore requires organizations to understand the data ecosystem surrounding their AI use — not simply the AI application itself.
5. Protect the trust of customers and employees.
This was another part of our conversation that I found particularly interesting because it moves the discussion beyond technology: AI has to be trusted to be adopted.
Aymen gave the example of remote service. It took years for some customers to become comfortable with the idea that a service provider could remotely access their machinery or systems and perform service.
Now consider asking a customer to trust an AI agent to conduct service. That's a very different threshold. And the same applies internally.
If employees don't trust the technology, they won't necessarily use it effectively. If customers have a poor experience with an AI capability, their confidence in the technology — and potentially in the organization providing it — can suffer.
This is why usability belongs in the AI defense conversation. So do accessibility, data privacy, and cybersecurity.
Aymen described these elements collectively as creating a kind of “middle layer of trust” in B2B relationships — a zone that helps establish confidence between the parties involved. This extends the goal from simply building AI capabilities that work to building capabilities that customers and employees are comfortable allowing to work.
6. Build AI literacy, not just AI capability.
Finally, I think there's a distinction emerging that organizations need to pay much more attention to: AI capability versus AI literacy.
There's a lot of focus right now on acquiring the people who can build AI, implement AI, and operate AI. Those capabilities are important.
But organizations also need people who can understand what is changing, connect those developments to business objectives, and determine what matters for their particular organization.
Because the pace of AI development creates a unique challenge. If there's another major development every day, it's impossible for an organization to chase every one of them. The answer can’t be to implement everything, but to develop enough understanding to make informed decisions about what deserves attention.
Aymen's perspective is that organizations that don't start building this knowledge now will face a much larger effort later.
And I think this is particularly important for service leaders. You don't necessarily need every service leader to become an AI expert. But you do need enough AI literacy throughout the organization to ask the right questions, understand the implications, and recognize where an opportunity aligns — or doesn't align — with the problems you're actually trying to solve.
A Balanced Approach to AI Innovation
As I said to Aymen during our conversation, I think it's easy to fall into one of two extremes with AI. On one side, there is the belief that AI is going to solve everything — including the skilled labor shortage — and that we should move as aggressively as possible.
On the other, there is fear of the unknown that can cause organizations to move too cautiously and miss meaningful opportunities. Neither extreme is particularly helpful.
What we need is the ability to hold both realities at the same time: There is enormous opportunity. And there is real risk.
Aymen's message wasn't to avoid the opportunity. His message was to adapt — with a strategy behind that adaptation. And that's where I keep coming back to the basketball analogy.
Offense gets the glory. It's the three-pointer everyone remembers. It's the new AI capability that gets announced, demonstrated, and celebrated.
But defense is what protects the lead. It means understanding where the vulnerabilities are, establishing the rules of engagement, and having people in place who can recognize when something changes.
In AI, that means defending the business, the information, privacy, intellectual property, customers, employees, and ultimately the trust that makes service relationships work.
We don't know everything that can happen next. We probably don't even know all of the questions we should be asking yet. But we can make sure we're building the awareness, governance, accountability, data protections, trust, and organizational capability needed to respond as the game evolves.
Because AI offense is getting the headlines. But a pairing strong offence with a strong AI defense is what will allow service organizations to keep playing.