The university is using conversational AI Virtual Support Agents (VSAs) to deflect 30–60% of incoming ticket volume before it reaches a technician.
Pitt IT’s Time to Resolution has improved by 40–90% through intelligent routing, AI-suggested responses, and automated status management.
As Pitt’s experience shows, TeamDynamix’s AI delivers 80%+ accuracy. This number improves as data quality investments compound.
Industry: Higher Education
Previous System: Salesforce
Ticket volumes are climbing, users conditioned by consumer-grade experiences expect near-instant resolution, and the labor market makes recruiting and retaining experienced IT support staff harder than ever.
For the University of Pittsburgh’s IT organization, supporting thousands of students, faculty, and staff across a large, distributed campus, this combination of pressures is all too familiar. The traditional answer of hiring more staff, working more hours, and more manual effort isn’t sustainable.
Pitt IT needed to work smarter. That meant putting AI and automation to work on routine requests and freeing the team to focus on the problems that actually require human expertise.
Tyler Puhatch, TeamDynamix Administrator at Pitt, works at the center of this effort, managing automation, AI workflows, and the ticketing systems that keep IT operations running. For his team, one of the most immediate and visible impacts of TeamDynamix has been what happens the moment a ticket lands in front of a technician.
TeamDynamix AI analyzes incoming ticket content and surfaces what the technician needs before they go looking for it: suggested field values, recommended routing, draft responses drawn from similar previously resolved tickets, and links to relevant knowledge base articles.
Pitt’s leadership specifically called out the suggested KB articles and similar tickets features as standout capabilities. The reason is practical: not everyone who handles tickets does so every day. From managers and specialists pulled in for specific issues, to part-time support staff, anyone who touches tickets occasionally benefits enormously from a system that meets them with context rather than expecting them to already have it.
“We all love something that’s going to make ticket work a little bit faster,” Puhatch said. “Obviously, tickets aren’t our job. Our job is to do the things that we’re reporting on in our tickets. We want to get back to actually doing real work.”
The onboarding impact is just as significant. New team members who would traditionally spend weeks absorbing institutional knowledge now receive contextual guidance on every ticket they open. What previously lived only in the heads of veteran staff is now available to everyone, from their first day.
Pitt was live with TeamDynamix AI ITSM quickly, following their initial boot camp. Their fast start reflects both the platform’s ease of implementation and the team’s preparation going in.
Beyond assisting technicians, TeamDynamix’s conversational AI Virtual Support Agent (VSA) is resolving a significant share of requests before they reach the service desk at all. Customers using the platform are deflecting 30 to 60 percent of incoming ticket volume, including routine requests handled end-to-end through automated, conversational interactions, with no technician involvement required.
For a university IT organization fielding the volume of requests Pitt does, that deflection rate isn’t a marginal improvement. It’s a structural shift in how capacity gets allocated.
Rather than patch an existing workflow that had accumulated years of complexity and manual friction, Pitt rebuilt change management entirely. “We were able to work together to get a new process for change management defined. We revamped the entire process, and we’re able to automate it end to end.”
When a ticket sits in “New” after a technician has already picked it up, customers assume nothing is happening and start following up. That follow-up creates noise that slows everyone down.
Pitt uses TeamDynamix’s integration and automation layer with webhooks to flip ticket statuses automatically the moment a technician is assigned — from “New” to “Assigned” or “In Process” — with no manual step required.
“Getting things to the right place quicker just gets a faster result for the customer,” Puhatch said, “and helps you get on to the rest of your day faster as well.” While this may seem like a small automation, it has had an outsized effect on customer confidence.
Documentation is consistently underdone in IT Service Management (ITSM), not because teams don’t value it, but because it always loses to the next ticket in the queue.
Pitt uses AI to close that gap.
By uploading TeamDynamix’s integration and automation application code and organizational templates to Claude, the team can generate formatted, accurate knowledge base articles in minutes rather than hours.
That documentation feeds directly back into TeamDynamix, improves conversational AI responses, and ensures the next technician who encounters the same issue has a faster path to resolution.
Pitt’s AI results didn’t happen by accident. They’re the product of a deliberate approach that starts well before any automation gets built.
Puhatch’s most consistent advice for organizations exploring AI ITSM is deceptively simple: start with your data.
“You want to keep your ticket hygiene as good as it can be. If you have things out there that aren’t correctly classified as an incident versus a service request, or have the wrong service collected, you want to get that cleaned up,” he explained.
Pitt backs that philosophy with automation.
The team has built workflows specifically designed to enforce ticket hygiene by validating field accuracy when tickets close and sending them back if something is missing or incorrect.
“We want what we’re ingesting into AI to be as tip-top as it can be,” he said. Cleaner data means more accurate AI outputs, and accuracy compounds over time.
TeamDynamix AI models are already delivering 80%+ accuracy even working with imperfect data, and that’s a number that only improves as data quality investments take hold.
Puhatch’s team works closely with departments across the IT organization, gathering pain points directly from the people experiencing them, then assessing impact and feasibility before anything gets built.
When the list of potential automation projects grows, and it always does once stakeholders understand what’s possible, impact wins, not ease, according to Puhatch.
One of Pitt’s most significant automation projects was a complete overhaul of its change management process. Rather than patch an existing workflow that had accumulated years of complexity and manual friction, Pitt rebuilt it entirely.
Puhatch was involved from the start as both technical builder and design partner. “I helped stakeholder ideas meet what is possible within TeamDynamix with the integration and automation layer,” he said. “We were able to work together to get a new process for change management defined. We revamped the entire process, and we’re able to automate it end to end.”
The result is a three-workflow architecture, each piece deliberately scoped for clarity and maintainability.
The first handles validation and routing — checking required fields, enforcing logical consistency, auto-populating standard changes from knowledge base articles.
The second drives the core change process, calculating risk scores, creating child tickets for the communications team when broader notification is needed, and routing approvals to the right service owners (with a built-in guardrail: a service owner cannot approve their own change).
The third manages scheduling and implementation, preventing conflicts and returning control to the submitter for execution and closure.
What took months to build now runs automatically. Having documentation and process mapping done upfront meant complexity was contained, not hidden, Puhatch explained.
Pitt’s approach offers a practical model for any IT organization navigating the same pressures.
The starting point isn’t the most sophisticated use case; it’s the foundation. Invest in ticket hygiene. Bring technical builders into process design conversations early so the gap between what stakeholders want and what the platform can deliver closes before anything is built. Let impact drive prioritization, not ease. Use reporting to establish a baseline, then track improvement after every automation goes live.
And don’t overlook the quick wins: knowledge base generation from existing tickets is a low-effort, high-value entry point that builds organizational familiarity with AI capabilities while improving the documentation everyone depends on.
The service delivery landscape has changed. AI and automation have moved from optional enhancements to operational necessities. The University of Pittsburgh is leading that shift, and TeamDynamix is the platform making it possible.
