Challenge
More than 4,000 tickets a month were triaged by hand. A three-person service delivery team worked a round-robin queue, correcting ticket types, agreements, and vague end user summaries before anything could be scheduled, and the backlog regularly violated SLA on distribution. When the board got busy, phones went unanswered.
Solution
TLC Tech deployed roughly seven CW AI Agents™ across the ticket lifecycle in ConnectWise PSA™: intake and triage; enrichment of summaries, agreements, and contact data; and board routing to the help desk. The NOC and backup boards are fully automated, including after-hours scheduling. The ConnectWise AI Assistant™ is enabled for every employee, and the agents integrate with documentation, CW PSA™, and ConnectWise RMM™.
Results
Tickets now reach a technician in about five minutes, down from 20-30 minutes. Average ticket time has fallen from a couple of hours to under one. CSAT is at its highest level to date. Calls are answered within one to two rings. After-hours and overtime have dropped, and the team has absorbed higher ticket volume without adding headcount.
From a round-robin board to an agent-first intake
TLC Tech is an MSP in the California market with 12 help desk technicians, four project technicians, and a three-person service delivery team, all reporting to a single service delivery manager. Rather than tiering its help desk, the company organizes technicians by role: sys admin, general application support, network administration, and SOC, with three technicians in each role.
That structure worked. The volume did not. More than 4,000 tickets arrived on the boards each month, and every one passed through three people first. A ConnectWise workflow rule round-robined incoming service requests to the triage team, who checked the type and subtype, confirmed the agreement was in place, and rewrote summaries so a future search or audit would surface the right ticket.
“Before, everything was manual. With more than 4,000 tickets a month, our triage team had to review every ticket, confirm the type and subtype, ensure the agreement was in place, and clean up the summary so the issue would be searchable later. End users are not technical experts. They might say the internet is down when a single website is unresponsive. That manual work took time, and it put pressure on our SLA for distributing service requests. That is what brought us to ConnectWise AI Agents.” - Hector Fanas, TLC Tech
Agents across the full ticket lifecycle
Today, CW AI Agents work from the start of a ticket’s life cycle. Everything is triaged by an agent, enriched with summary lines, agreement, and contact information, and then handed to the board agent, who passes it to the help desk, where dispatchers assign it to the best-fit role. Fanas estimates that about seven agents now run across the service. On the NOC and backup boards, the loop is closed entirely: those tickets are triaged, scheduled, and dispatched without a person in the middle.
That matters most outside business hours. CW RMM™ alerts that hit the ConnectWise NOC Services™ or ConnectWise SOC Services™ board are now triaged and scheduled directly to the after-hours technician, who starts a shift with work already queued rather than hunting for it.
“After hours, our RMM alerts go to the NOC or SOC board. Now those tickets are automatically triaged and scheduled with our after-hours technician. They do not have to go looking for the work. It is already queued for their night and early morning.” - Hector Fanas, TLC Tech
20-30 minutes down to five
TLC Tech tracks its service metrics in ConnectWise Reports and Dashboards, and the numbers have moved quickly. Before CW AI Agents, a ticket sat 20-30 minutes before it was triaged and assigned, closer to 30 minutes on a Monday. Now the movement between boards, from arrival to being processed for a technician, is five minutes or less.
Average ticket time followed. Five to six months in, a metric that used to take a couple of hours now takes less than one hour. Fanas attributes that to both faster dispatch and the CW AI Assistant™, which does a first pass of research before a technician opens the ticket, pulling history, surfacing lookalike tickets, and flagging blind spots worth checking.
The same assistance extends to administrative work. Fanas said TLC Tech has enabled the tool for everyone, including his own billing workflow, where the agent reduces the need to search through large volumes of tickets to confirm agreements and exceptions before invoicing.
We used to measure average ticket time in hours. Now we are below one hour. [CW] AI Agents help on the triage and dispatch side, and the [CW] AI Assistant helps before the ticket reaches the technician by pulling history, researching similar tickets, and surfacing context they can use right away.
Time back to answer the phone
The three people who used to clear the board did not get automated away. They got their day back. TLC Tech’s tagline is “we speak human,” and the dispatch team now has the capacity to answer every incoming call live, with no rollover to the service desk manager. Clients noticed the change before anyone told them about it.
On busy days, clients used to notice that the phone would keep ringing until it rolled over to overflow. Now we answer within one to two rings. Clients ask whether we hired more people, and that change has helped bring our CSAT higher than ever.
There is a second-order effect on resolution quality. Without the pressure to get callers off the line and back to the board, the team can ask qualifying questions, such as “Is this happening to you, to a few people, or company-wide?”, which routes the ticket to the right technician the first time and makes a one-touch close far more likely.
A technician who started at tier zero
The strongest example of leverage at TLC Tech is a technician the company hired in December who had no IT background at all: a former delivery driver with a psychology degree and no industry experience. He is on track to move up to Tier 2 sys admin by the end of the year.
TLC Tech has been in business since 2004, and some customer relationships date back 15 years. Reconstructing that history used to be the hardest part of being new. With the CW AI Assistant, a technician can simply ask for it.
“Imagine being brand new and needing to understand a customer we have supported for 15 years. Now you can ask the [CW] AI Assistant for the history of a computer and whether there have been issues in the past 60 or 90 days. That helped him track down the root cause faster. He did not have years of technical history yet, but he had the communication skills to use the assistant and get to a resolution much faster than he would have manually.” - Hector Fanas, TLC Tech
The same effect appears across the help desk. TLC Tech gamifies ticket closure with a leaderboard; on one observed morning, technicians were already ahead of the team’s five-KPI daily ticket target within the first hours of the shift. Because routine work, such as password resets, printers, and small recurring tickets, can be handled by agents or converted into one-click scripts, technicians spend more time on higher-level problems. Activity in the company’s internal tech-questions channel has gone up, not down: with less pressure, people have time to ask, answer, and level each other up.
The documentation flywheel
TLC Tech entered the project data-rich. The company has used documentation since its inception, so there was a strong knowledge base to guide agents. CW AI Agents were integrated with documentation, CW PSA, and CW RMM. Material that had been sitting in SharePoint before the documentation era was also uploaded and used to generate SOPs that reflect how TLC Tech actually works.
The loop now runs in both directions. When a ticket is marked complete, a workflow generates an SOP from the resolution notes. A person reviews it for accuracy, and it goes back into documentation, so the documentation agents’ read gets better every time a technician closes a ticket.
“As soon as a ticket is marked complete, our workflow updates the documentation and generates an SOP from the resolution notes. A person reviews it for accuracy, and then it goes back into our documentation.” - Hector Fanas, TLC Tech
Retiring a homegrown AI for something native
TLC Tech was not new to AI. The company had built and run its own model for roughly two years before adopting CW AI Agents. Maintaining it was the problem: a small team running on limited hardware had to spend weeks feeding data, managing swaps and maintenance windows, and using connectors to integrate AI into the tools where work was already happening.
We built our own AI and used it for about two years, but ConnectWise brought that capability natively into the platform. We no longer have to connect everything through Zapier or manage the data feeds ourselves. New data from the week is fed over the weekend and ready for the following week. It means fewer headaches for me and lower costs for the business, and we can reinvest that money into the company and our people.
Looking ahead
The business case has expanded beyond the service desk. Fanas reports that after-hours work and overtime have dropped, and because the day now has capacity, TLC Tech has been able to take on more tickets without adding another person to the payroll. For him, the goal is not fewer technicians, but technicians focused on work worth doing.
“One thing I tell my technicians is that I want to pay them to do nothing by the end of the day. As an engineer, do you want to spend every day changing passwords or fixing printers? With [CW] AI Agents and the [CW] AI Assistant, they can move past more of that routine work and focus on higher-level tickets where they learn more.” - Hector Fanas, TLC Tech
Key takeaways
- Agent-first intake compresses the distribution window: TLC Tech reports that the time to triage and route a ticket fell from 20-30 minutes to about five, easing the SLA pressure created by manual dispatch.
- Resolution gains can come from context as well as routing: Average ticket time dropped from a couple of hours to under one hour within roughly six months, helped by faster dispatch and assistant-led research before work begins.
- Autonomy matters most where coverage is thin: More than 4,000 tickets a month now start with AI triage, while NOC and backup workflows can schedule after-hours work without a dispatcher.
- Capacity returned to the customer relationship: A three-person dispatch team shifted from clearing the board to live-answering customer calls within one to two rings, and TLC Tech reports CSAT is at its highest level to date.
- AI can extend existing team judgment: A technician hired with zero IT experience is on track for Tier 2 in under a year, using the AI assistant to reach customer history and root cause that he could not have reconstructed manually.
- Documentation improves when closed work feeds the knowledge base: Closed tickets automatically generate SOPs that a human reviews and publishes to documentation, so the knowledge base the agents read improves with every resolution.
- Native integration can replace brittle AI maintenance: Replacing a self-built, self-maintained AI with native in-platform agents removes integration and maintenance overhead and lowers cost.
- The business case reaches beyond ticket handling: TLC Tech reports lower after-hours work and overtime while absorbing higher ticket volume without adding headcount.
