How Truemed Hit 70% AI Resolution and Cut TTR by 93% in 90 Days with Applied Labs
Highlights:
70% AI Resolution Rate up from ~30% with a previous AI tool, with meaningfully higher response quality, achieved in just 90 days.
93% reduction in TTR (Time to Resolution) from weeks to under a day, maintained through peak season volume spikes.
“I don't think you get this level of partnership with any other customer support tool.”
Workflow complexity that the team never dreamed was possible including live database lookups that used to require a full-time employee.
CX repositioned as a revenue and product function, not a cost center, with weekly intent audits directly influencing the product roadmap.
Company:
Truemed, a venture-backed HSA and FSA marketplace that has raised from top investors including Andreessen Horowitz, sits at a rare intersection, part telemedicine platform, part payment processor for over 3,000 brands. With multiple deployment models, denied-claims workflows, and enterprise merchant partners like Whoop and Peloton, their support operation is anything but standard.
When rapid growth pushed their legacy stack past its limits, they needed more than a helpdesk upgrade. They needed a platform that could match the complexity of their business.
The Pain:
Truemed had been running on Front for ticketing and Lindy for AI automation. On paper, the setup worked until it didn't. The business doubled between October and December 2025, and the volume also doubled overnight
The team had built out basic AI flows in Lindy that achieved roughly 30% resolution, but the underlying approach was fragile. There was no meaningful quality evaluation, no analytics layer, and no path to connecting the AI to Truemed's own systems.
On a scale of 0–10, Colin rated the pre-transition pain at an 8 or 9.
The core issues were structural:
- Two disconnected systems (Front + Lindy) made troubleshooting AI responses slow and opaque, with no unified view of what was happening.
- No connection to internal data. With 90% of Truemed's answer logic living in their own database, the AI couldn't access what it needed to respond accurately.
- No visibility into customer intent. Manual tagging in Front was inconsistent and unsustainable.
- Volume that couldn't scale. A human team can only answer so many emails. As Colin put it: "A human can't answer more than 200 emails. That was it."
What Truemed Evaluated and Where Others Fell Short:
Truemed evaluated several AI vendors, including Decagon and Sierra, before choosing Applied Labs. The common theme across those conversations was the same: platforms built for standard use cases struggled to accommodate the nuance of Truemed's business.
For a company with healthcare-adjacent workflows, claims logic, different deployment models, and no standard Shopify infrastructure, most vendors offered rigid, template-driven solutions that couldn't flex to fit. Feature requests that were table stakes for Truemed's operation were either unsupported or required significant enterprise commitments to even discuss. The gap between what existing tools could do and what Truemed actually needed was too wide to bridge.
Why Applied Labs:
Truemed chose Applied Labs for three reasons: technical depth, speed, and genuine partnership.
Database-first architecture. The ability to connect directly to Truemed’s data was the deciding factor for Tiffany Kim, Truemed's COO. Within days of that conversation, the integration was ready. For Colin, what followed still feels extraordinary:
Workflow complexity that wasn't possible before. Applied Labs didn't just automate FAQs, it unlocked flows the team had never even considered building. Multi-variable eligibility lookups. Merchant-specific response synthesis across multiple deployment models with no one-to-one knowledge articles.
A team that ramps fast and delivers faster. When Truemed flagged issues or made requests, the turnaround was measured in hours. The Applied Labs team got high context on Truemed's business quickly and produced work product that exceeded expectations from day one.
Gerard Mealy, Product Operations Manager, echoed the same sentiment about the engineering team's responsiveness:
Results:
70% AI Resolution Rate with Accuracy that Actually Holds Up
Truemed more than doubled their AI resolution rate in 90 days while dramatically improving the quality of responses. The distinction matters: the 30% Lindy achieved came with lower-fidelity answers and lower visibility into errors. Gerard does weekly quality audits and describes the difference clearly:
The agent's ability to synthesize across multiple knowledge sources, rather than finding the single most relevant article and rewriting it, is what enables accuracy at this complexity level.
93% Reduction in TTR: From Weeks to Under a Day
Mean time to resolution is now under one day, a 93% reduction maintained even as ticket volume grew.
More importantly, the team now feels prepared rather than anxious about the next peak season. Today it's a different story entirely.
A Unified CRM Where AI Is a 24/7 Colleague
One of the most profound shifts has been in how the team experiences their day-to-day work. Unlike legacy platforms where human agents have to sort through what the AI did or didn't handle, Applied Labs operates as a true unified system. Human agents now log directly into the escalated queue because the AI agent has already taken care of everything else.
The trust built in the AI's responses is what makes this model work. As Gerard put it: the confidence in Applied Labs' accuracy means human agents aren't second-guessing the queue. They're just focused on the work that actually requires them.
With that bandwidth recovered, the team now spends time on merchant team conversations, product roadmap analysis, and systems thinking that was previously impossible. A mobile UX bug caught in the inbox this week. None of it would have been visible in the old model.
From Cost Center to Growth Driver:
Truemed's leadership had long believed CX was a user research function, not a cost center. Applied Labs made that vision possible:
The team runs weekly audits of topics and intents, tying patterns in the inbox directly to product decisions. Bugs that would have been invisible in aggregate analytics surface clearly when support data is properly classified.
The team is also rethinking headcount strategy entirely. Rather than scaling linearly with volume, the goal is to keep the team lean while the AI handles the repeatable tier — and deploy human expertise where relationships and judgment actually matter.
Tiffany summed up the broader strategic shift: in Q4, Gerard moved from product management into product ops precisely because the team had decided support wasn't just a service function it was a revenue-generating one. Applied Labs is what made that bet pay off.
What's Next:
The team has set an internal goal of reaching 85% AI resolution rate. The Merchant team is next to adopt the platform at the depth the customer team already has. And Tiffany is now tracking tickets-per-user as a north star KPI, a metric that was previously onerous to measure.
Get in Touch:
If you want to automate complex support, connect AI to your internal systems, and turn CX into a strategic growth function, Applied Labs can help.
Subscribe to Applied Labs' blog
Get notified about new product features, customer updates, and more.
Related posts
Introducing Applied Assistant
Applied Assistant is a managed AI assistant for company work. Ask in plain language, connect the tools your team already uses, and let it learn your tone, your policies, and your customers. The more you work together, the better it gets.
An Outcome Is Only as Honest as Its Definition
Outcome-based pricing is where AI is taking the category — but the word 'outcome' now means wildly different things depending on who's selling it. What we believe, and why the definition is the whole game.
Applied Labs is now generally available
Applied Labs is generally available with public pricing, a 14-day free trial, and the full platform teams need to run AI agents, human support, helpdesk, CRM, analytics, and governed CX operations in one place.