When ChatGPT debuted and generative AI took off, we were brainstorming new ways for ServiceMax to help with knowledge transfer from retiring field service technicians. It’s one of the top challenges our customers face, and traditional methods like manuals and knowledge management systems weren’t cutting it. So, the use of large language models to train a system on a data set was eye opening. We knew right away this approach could work for ServiceMax too. Fast forward several months and a customer beta program later, and we’re bringing you ServiceMax AI. ServiceMax AI has access to all the data stored in a ServiceMax instance, engages with technicians through natural language and an easy-to-use chat interface, and completes tasks through specialized AI agents.
Our goal was to create an experience that feels like you’re working alongside another human technician, dispatcher, or operator. A technician can type or ask a question, and the system goes to work with the agents behind the scenes. We focused on three agents to start:
- Work History: Understands work orders, replaced parts, and service history.
- Scheduling: Manages scheduling, upcoming appointments, and modifications.
- Knowledge Access: Retrieves information from PDFs, manuals, and unstructured documents.
The magic happens when these agents are combined. ServiceMax AI dynamically selects the most relevant agents to answer a given question, just as an experienced technician might tap into their memory and knowledge to troubleshoot an issue. What really sets this apart is how the agents work together behind the scenes, supported by today’s AI techniques, like retrieval-augmented generation (RAG). We built an entire orchestration and reasoning layer that uses RAG to interact with agents to get the answers it needs to provide a response to the overall question.
Traditionally, software accesses data in a deterministic way—using predefined APIs that only work for specific queries. This rigid approach can be limiting when the question is unexpected or nuanced.
So, we built ServiceMax AI to be able to access data in a non-deterministic way. This means the AI doesn’t rely on rigid API structures. Instead of predefining every possible method for retrieving data, we’ve "taught" the system how to access and retrieve information dynamically within ServiceMax. It mimics human thought processes—recalling experiences, analyzing context, and providing insightful answers in real-time.
It's like the old adage about teaching someone to fish versus giving them a fish. That’s essentially what AI technology enables us to do, and by establishing a foundation for a more intuitive and efficient way to access the right data, we unlock the true potential of field service knowledge for our customers.
When we consider the day-to-day reality of a field technician, it’s clear that chat is the ideal interface. Chat is powerful because it’s consistent with how most people already communicate. Technicians are always on the go—whether in their trucks, on job sites, or in the field. They don’t have time to navigate complex systems or sift through lengthy documents. They need helpful information quickly, delivered in a manner they’re comfortable with.
ServiceMax AI Chat provides exactly that: a natural, efficient, and familiar way for technicians to interact with ServiceMax AI. Just as they use ChatGPT and other AI tools in their personal lives, they can now leverage the same intuitive chat experience in their professional work with a system that’s designed exclusively for field service. With just a few typed questions, they can access critical information—just like reaching out to a colleague.
We're excited to bring ServiceMax AI to the field and see how it evolves. After months of beta testing with our customers, we’re encouraged by how it fits into a technician's workflow—but this is just the beginning. AI is advancing rapidly, and as more customers adopt it, we’ll continue refining and improving ServiceMax AI based on real-world feedback.
We hope you have a chance to test out ServiceMax AI. Feel free to share your feedback or comments with me on LinkedIn, where I’ll be sharing more of my thoughts around ServiceMax and AI. If you're seeing other interesting ways AI is impacting field service, I’d love to hear about it.