By Sarah Nicastro, Founder and Editor in Chief, Future of Field Service
It’s become evident to me that there are some truly mixed feelings among service leaders when it comes to AI.
On one hand, I’ve heard plenty of “I’m so sick of hearing about AI.”
On the other, in our most recent Stand Out Service Trends report, 90% of leaders say that advanced AI is crucial for the future of service.
I can understand feeling torn; in many ways, I do too. I’m pro-human in every way, so I too tire from all the AI talk – and have my concerns around the potential implications it could have if not harnessed responsibly.
But it’s also become undeniable how – in the right instances – AI is unmatched in how it can help alleviate some of the pressure of the workforce gap, ease burdens for the frontline, and allow scalability of service outcomes that humans alone cannot.
The ideal human/AI partnership was on full display at IFS’s recent Unleashed event in Orlando, as I described in my article last week. Seeing the examples come to life at the event was really helpful in understanding the real-world context of where AI can be most beneficial in service-centric businesses.
Let’s explore five scenarios where AI, particularly agentic, can have a positive impact on service operations and outcomes:
#1: The Parts Agent
A technician is scheduled to repair a compressor, but the required part isn't in their van. Instead of making the technician search multiple systems or call the parts department, an agent identifies the correct component for the specific asset, checks inventory locations, and initiates procurement if necessary.
This is real-world example shared by Kodiak Gas Services at Unleashed. Pedro Buhigas, CIO, explained how Kodiak's material replenishment agent has improved the efficiency of this process by 85%.
What makes it agentic? It doesn't stop at recommending a part. It coordinates the steps required to get it where it needs to be.
Potential next step: Have the agent coordinate delivery to the technician's next location, while flagging substitutions or high-cost purchases for approval.
#2: The Expertise Agent
One of the biggest points of potential for AI in field service? Giving every technician access to your best engineer’s knowledge.
A technician encounters an unfamiliar fault. An agent searches technical documentation, past repairs, parts information, and equipment history, then guides the technician through the most relevant troubleshooting steps. If the issue remains unresolved, it escalates with the diagnostic information already assembled.
This process could significantly reduce phone calls, messages, and drain on the time (and energy) of the organization’s experts, as well as save time on manually searching for information, speeding repairs and eliminating frustration.
What makes it agentic? The system helps progress a repair through multiple steps, rather than merely retrieving a document. Potential next step: Let the agent recognize when it has reached the limits of its knowledge and bring in the right human expert with a concise summary of the problem and steps already attempted.
#3: The Incident Agent
A connected device indicates that a fault may be developing. An agent correlates alarms, historical incidents, and customer reports; determines whether remote remediation is viable; and coordinates a field dispatch only if one is necessary.
One such example is AT&T, which has reported that its AI-enabled end-to-end incident management system, which added AI agents in 2025, helped prevent 3.1 million unnecessary field dispatches and reduce customer downtime by more than 12 million hours over the preceding year. (The overall results reflect the broader AI-enabled system, not agents alone.)
What makes it agentic? It moves from identifying a problem toward coordinating resolution, including deciding when a human technician is needed.
#4: The Job-Completion Agent
Imagine a technician finishes a two-hour repair and needs to document the fault, work performed, parts consumed, readings, and follow-up recommendations.
The agent interprets voice notes, photographs, equipment readings, and work-order details. It drafts the service report, reconciles parts used against inventory, flags missing information, and prepares the job for invoicing. Less administration for technicians, faster invoicing, cleaner asset histories, and better data for future AI applications.
The story shared by Mo Dawood, CIO of Eight Group at IFS Unleashed offers an example of a real-world starting point: its use of image analysis reduced documentation time from 25 minutes to one.
What makes it agentic? It coordinates the completion workflow across field documentation, inventory, and billing rather than merely transcribing notes.
#5: The Service Growth Agent
A technician repeatedly finds that a customer's equipment is operating outside recommended parameters or discovers that an aging asset would benefit from an upgrade.
With the appropriate permissions, an agent compares the findings with the customer's installed base, service agreement, maintenance history, and available offerings. It creates a qualified opportunity, prepares a tailored recommendation, and routes it to the account manager — without asking the technician to become a salesperson.
What makes it agentic? It connects operational evidence to a commercial action and coordinates the handoff.
This could lead to more relevant service-contract renewals, upgrades, parts opportunities, and outcome-based offerings — with less dependence on individual technicians remembering to record every opportunity.
This example also connects to a larger point for service: AI has potential in service beyond simply improving efficiency or reducing costs; it can help service organizations grow revenue and differentiate through outcomes.
The Bigger Story: Looking Beyond Isolated Agents to Coordinated Work
While it can be helpful (and a great crawl/walk/run starting point) to consider these specific examples, they become even more interesting when we look at what becomes possible when they work together.
Imagine an industrial asset showing signs of failure. One agent detects the issue; another evaluates its severity; a third identifies the parts and technician needed; another coordinates scheduling and customer communication; and a final agent assembles the service report, captures the findings, and identifies follow-up opportunities.
The value isn't five separate AI features; it's a coordinated process that moves from detecting a problem to resolving it and learning from it, with humans stepping in where judgment, safety, customer relationships, or exceptions demand it.
Moreover, the opportunity isn't to put an AI agent into every corner of field service. It's to identify the points where work gets stuck — where people are waiting for information, chasing parts, coordinating schedules, or translating one system's output into another system's action. Those are the moments where agentic AI can begin to change not just how efficiently service is delivered, but what service organizations are capable of delivering in the first place.
An Important Reminder
To bring it back to needing to strike right balance of leveraging AI’s value without overinflating its role, I wanted to share this reminder that Gyner Ozgul, CEO of Fortis Fire & Safety posted on LinkedIn last week:
“Three out of four AI users now ask a chatbot before they ask a colleague. If you value the ability to network with peers, read on.
I get it. I use AI every day. It’s fast, it’s always available, and it never makes you feel like you’re imposing. But I keep thinking about how I learned before any of this existed.
Years ago, I was an operations and service guy who got asked to lead a supply chain role. I knew nothing about supply chain. There was no AI to lean on. So, I reached out cold on LinkedIn to leaders across the supply chain I was now overseeing. People I’d never met. I asked a lot of basic questions. Those conversations turned into the solutions we built, and a lot of them turned into real relationships, ones I value even today.
AI would have given me answers faster. It wouldn’t have given me those people. We watched social media do this to generations of kids. Real connection got swapped for something that felt close enough. I think a lot of professionals are walking into the same trap without noticing. I’m not saying stop using AI. I’m saying use it to get ready for the conversation, not to skip it.
Let AI handle the basics, then call someone with a sharper question. Pick one person a week you haven’t talked to in six months, or one you’ve never met. Send them something useful before you ask for anything. And if you lead a team, say ‘I don’t know, what do you think?’ out loud. It gives everyone else permission to ask too. Plenty of people tell me they skip the call because they don’t want to bother anyone. Bother them. Most people want to be asked.”
As AI use cases in service expand, we need to be diligent in considering how we continue to prioritize humanity. How we nurture relationships, avoid isolation, and how we protect the brainpower required for critical thinking as we come to embrace new information-rich processes. So, go crazy in envisioning how AI can be an asset to your service operation – but spend equal time and energy ensuring you protect the people aspects that make it special.