Complex triage work consumed its best people

"We were using our best people to prepare cases instead of solve them," says Mark Britnell, C Spire's Senior Manager of ServiceNow Platform.

C Spire stands for 'Customer Inspired'—there's a stick figure in every meeting room, on every screen, representing the customer at the center of every decision. That philosophy made the triage problem impossible to ignore: the support team's best talent was disappearing into preparation instead of resolution.

Two support agents handled triage: verifying customer contacts, selecting service categories, flagging outages affecting multiple accounts, routing cases. All day. About six thousand total cases per month, including roughly three thousand emails.

The work was complex. A single email might belong to multiple customer accounts. Location mattered. Outages affected multiple customers at once. And each decision fed into reporting and compliance.  With information spanning multiple tools and channels, staying aligned across teams was more challenging.

Traditional automation couldn’t handle domain logic

C Spire's support structure required coordination across multiple systems and channels. When a customer's location changed or an outage affected multiple accounts, getting that intelligence to the right team took time and manual follow-up.

The resolution team never had the full context they needed. And the triage team, highly trained support staff with years of telecom domain knowledge, spent their days on coordination instead of resolving issues.

The problem wasn't just volume, it was also complexity. A single AI agent might verify a customer contact correctly. But if that contact works for multiple accounts, and the agent only sees one account record, it classifies the case wrong. Or it detects a potential duplicate but doesn't have location context to know if the duplicate is actually a separate outage. Or it flags an incident as major, but the routing logic doesn't understand account hierarchies, so it goes to the wrong team anyway.

But single-agent automation breaks when each decision depends on the previous one. A single agent can't understand that verifying the right account has to come before classification, which has to come before routing. C Spire needed something different: multiple agents that could hand off context to each other and preserve domain knowledge across the entire workflow.

Three AI agents automated the triage workflow

Britnell's team started building ServiceNow AI Agents within their Telecommunications Service Management (TSM) system—the platform C Spire had been running since 2022. But it wasn't just the system itself that mattered. It was the data foundation: four years of telecom-specific service data, account hierarchies that understood multi-customer relationships, location schemas for coverage areas, and years of outage history. TSM was the knowledge base that made the domain logic possible.

That's exactly where ServiceNow's agentic AI made the difference: orchestrating, not just automating, multiple agents that understood their domain logic based on four years of telecom-specific data securely managed within their system. Domain-aware orchestration.

The goal was to encode the business logic that their triage team had been manually executing: three AI agents working together in one agentic workflow:

• First agent: Verify customer contact. Handle cases where one contact works for multiple customers.
• Second agent: Classify the case by service offering, location, category, and subcategory. Feed that classification into reporting.
• Third agent: Detect potential duplicate cases and surface possible major incidents for human verification.

"I was a little skeptical," Britnell says. "Our data is complex. Our configuration is unique. I told our team, I'd be shocked if this works."

Within six months of starting the project, the agents went live.

Today, about 70-72% of inbound email cases move through triage and routing automatically. The coordination bottleneck nearly disappeared.

The team stopped preparing and started solving

When cases arrive pre-classified and routed with full context, the resolution team moves straight to diagnosis.

A customer password reset that used to get escalated because there wasn't time? Now it gets solved immediately. A location-specific service issue? Routed with location context already attached. Triage handling time dropped 44%, from 54 minutes to 30 minutes. More importantly, case closure improved by 125%.

"With AI agents handling triage, everybody can do the same thing at the same time—we got that time back, and it flows straight to customers." Britnell says.

C Spire's support leadership felt it immediately. When Britnell shared the first month's metrics, the response was immediate. The organization celebrated the win internally.

The team got back more than 283 hours per month from manual triage. That capacity went straight back to customers.

"Triaging emails was necessary," Britnell explains. "But it was tedious. Now, with support from AI agents, the team can focus on what they were hired to do: resolve customer issues."

Next up: C Spire will scale orchestration across incident and order workflows

This deployment validated something bigger than triage. C Spire doesn't just run ServiceNow internally, it sells managed services built on the platform to other customers. “It wasn't just an internal win; it proved that agentic AI can operate at the level of real-world domain complexity,” added Britnell.

When orchestration works this well, 70-72% email triage automation, 125% case closure improvement, and 44% triage reduction, the roadmap accelerates. Mark Britnell is already scoping AI agents for incidents and orders.

That's what "Customer Inspired" looks like in practice. The stick figure in the meeting room—the customer at the center of every decision—is truly at the center of what the support team does every day.