The Background
Elevating trust through better support
For an aspiring personal trainer halfway through an ISSA certification, support isn't a nice-to-have — it's the difference between finishing the program and walking away from a career change.
ISSA's students span time zones from California to Cairo. Many are studying after their day jobs. Many have financially invested in a credential that will determine whether they can work in the fitness industry.
When they have a question about coursework, payments, or their final exam, they need an answer fast. The brand promise of we make learning convenient, effective, and set you up for success from day one hinges on it.
But that promise was slipping. Students were waiting too long for an email response during business hours, and up to 40 hours if they submitted their request after phone calls went unanswered.
For a student trying to time a final exam around a work shift, or a Spanish-speaking trainer in São Paulo trying to confirm an enrollment on a Sunday night, that gap wasn't an inconvenience — it was a quiet message that nobody was on the other end of the line, at exactly the moment they needed someone to be. Trust in the brand was being eroded.
The Challenge
A New Leader, an Inherited Operation, and a Volume Problem
Ashley Thompson, ISSA's Director of Support, inherited the operation. What she walked into was a support team with processes that had not been scaled even though inbound traffic was up to 20,000+ per month.
ISSA was running on Freshdesk, used not just for support but as the company's entire communication platform. "Everyone owned it, so no one did," Ashley explained. Reporting was opaque. As the leader accountable for response times, she couldn't cleanly answer questions about her own team's performance.
The volume was roughly a hundred times what she'd handled at her previous company, and ISSA wasn't staffed for 24/7 coverage. Headcount alone wasn't going to close the gap. The math didn't work, and the abandonment rate had been stuck at an unacceptable rate for months.
The team was burning out. First response time had been flagged as a top executive priority. The CEO and executive leadership were leaning into AI as a strategic direction. Ashley had a window, but no easy lever.
The Solution
Building an AI Teammate Students Actually Trust
What students needed was simple: an answer in the moment they asked, in their language, on their schedule. What Ashley needed was a way to deliver that without sacrificing the quality her team had spent years protecting. She moved support out of Freshdesk into HubSpot Service Hub, then turned her attention to the AI piece.
The First Launch That Failed, and the Unlock That Followed
Ashley's first attempt to deploy Customer Agent was a stumble. Within days of going live, she pulled the plug. The answers weren't good enough. But instead of blaming the AI, she diagnosed the real problem: ISSA's knowledge base — 250 articles built over years for human readers — wasn't structured for an AI to retrieve from cleanly.
So she rebuilt it for AI. She fed six months of ticket data, page analytics, and search behavior into an AI model and asked it to redesign the help center for AI consumption. The result was 35 articles, a 7x compression, with no negative impact on the 10,000+ visitors a month who relied on it.
"That was the unlock," Ashley said. "The AI was never the problem. The knowledge base was." It's the lesson she now tells anyone deploying Customer Agent: the agent is only as good as the data you feed it.
Meet Tessa
With the knowledge base rebuilt, Ashley turned HubSpot Customer Agent back on — for chat first, then for email tickets. The team named her Tessa. She runs on the public help center for general questions, and on a private "back-door" knowledge layer Ashley built for tier-1.5 issues: the conditional answers that shouldn't be advertised on a public page but should be given to a student who asks. Tessa is also bilingual.
With the knowledge base translated into Spanish, she serves ISSA's South American students in their own language without any separate setup.
Then came the wow moment. Tessa was answering students in under 70 seconds, often before ISSA's own auto-responder fired. Some students wrote in surprised: I just sent this and it's already answered? For half of them, that was the best part of the experience.
The other half just couldn't believe an AI agent answered them so well, and so kindly. The data proved it; Tessa's service rating matched the team average.
From Tessa, Forward
A migration to the Help Desk workspace compounded the gains. Ashley's top human agent went from 100 tickets a day to 175, with fewer clicks, response recommendations, and automatic ticket categorization. Tessa absorbed routine email volume; her teammates handled the complex, emotional, high-stakes cases where human judgment matters most.
The first chat run after the relaunch told the same story at scale: monthly chat volume jumped from around 200 conversations to 4,000+, with a single-digit escalation rate.
The Transformation
From "Drowning" to Expanding What Support Can Do
The numbers moved fast. Business-hours first response dropped by more than 50%; more than 70% of students started getting a response within 1 hour. After-hours response dropped so dramatically students believed the team was live 24/7.
AI handling of total inbounds climbed from 15% in August to 40% by December, across chats, email tickets, and calls. Tessa now addresses around 50% of inbound email tickets fully and handles 100% of chats with human handoff. Her CSAT currently runs slightly above the team average.
But the metric Ashley cares about most is the one she didn't expect to fix: phone abandon rate. With Tessa absorbing email volume, Ashley redeployed agents from email to phones.
The abandon rate fell to just 15%. Students who used to give up before reaching anyone were finally getting through. The brand promise of we're here to help started landing again, in real time, in two languages, around the clock — all while addressing the underlying concern of team burnout.
A Team That Stopped Drowning
The harder transformation was inside the team. When Ashley first introduced AI, her team was scared, convinced this was a layoff dressed up as innovation. ISSA's leadership, like most leadership teams, asked the predictable question: can this reduce headcount? Ashley pushed back hard.
She framed Tessa to her team as "a tool to wield, not a sword to fall on," and made the case to executives that the real opportunity wasn't fewer humans. It was more service. Proactive outreach. Better coverage. Capacity for the work the team had never had room to do. Eliminating headcount is not the primary outcome of a tool like Tessa.
That reframe held. Reporting that used to take days now takes five minutes. Ashley ships changes (autoresponder edits, routing rules, knowledge updates) herself, without waiting on dev or admin tickets.
In student feedback, one of the highest-rated members of the support team turned out to be Tessa; students praised her by name, sometimes not knowing she wasn't human.
The most telling change, though, wasn't on a dashboard. Tessa absorbed nearly 4,000 conversations in a single month without missing a kid's baseball game, a school pickup, or a Friday night off. The team got those instead.
People who'd felt perpetually behind, working through evenings and weekends just to keep up, got their lives back. The work got better because the humans no longer felt behind every day when they signed off.
For Ashley, the deepest shift is one of posture. Her team has gone from perpetually overwhelmed to actually present for the students who need them most. The fear has been replaced by something she didn't think was possible six months ago: confidence in what comes next, and pride in what they've already built.
.png?width=215&height=50&name=Ignite%20Logo%20(1).png)
%20(1).png?width=215&height=50&name=Fusion_Logo_Vert_4c_RGB%20(2)%20(1).png){kind=logo}
-1.png?width=215&height=50&name=tka-purple%20(2)-1.png)
