How TrueCX Uses AI to Transform Contact Center Agent Training and Validation
I've spent 20+ years in contact centers: Sprint, nearly a decade at NICE, CRO at Balto. Everywhere, agents learned on live customers because there was no better way to practice. I founded TrueCX in 2024 to give them one.
By Lonnie Johnston
· Career contact center guy: analytics at Sprint, a decade at NICE, CRO at Balto, now founder.
· Fairview, TX
Published June 16, 2026 · 6 min read
Published June 16, 2026 · 6 min read
This case study is based on responses submitted directly by the founder or member of the team from TrueCX. They have verified ownership of their domain truecx.com on SaaS Browser.
How TrueCX got started
The moment I keep coming back to is a new agent sitting with her headset off, trying to collect herself after her first week on the phones. She had passed training. She knew the policies. None of it prepared her for real customers.
I'd spent over 20 years in contact center technology by then, at Sprint, NICE, and Balto, and I'd watched that same scene play out the whole time. We hand agents scripts and quizzes, sit them next to a tenured rep for a few calls, then turn them loose and let actual customers be the practice. The first angry caller, the complex claim, the person who won't accept a scripted answer - that's where they learn, live, on someone real. The training passes, the agent doesn't.
That's what I started TrueCX to fix in 2024. We built AI customers that sound, react, and object like real ones, so agents get realistic practice before the stakes are real. They can have the hard conversation ten times in a morning and never burn a real customer doing it. The same technology now validates AI agents too, but it started with her.
Growing TrueCX: what worked and what didn't
What worked is almost embarrassingly old-fashioned, happy customers and warm introductions. Our best pipeline has come from customers getting real results and talking about them, and from a handful of industry connectors who will put their name behind an intro. When one of our healthcare clients cut handle time 20%, that story did more for us than any campaign. A referred prospect shows up already half-convinced. A cold one shows up looking for reasons to say no.
The flop was classic cold outreach. Cold email and cold calling have gotten brutally hard, and AI made it worse, not easier. Everyone's inbox is flooded with machine-written messages, so buyers now assume anything unexpected is a bot and delete it in two seconds. The irony is that people crave human connection more than ever. They just reject it on sight because they've been burned too many times. The lesson for us, you can't automate your way to trust. You have to earn it through customers, proof, and real relationships, then let outbound support that instead of leading with it.
What TrueCX customers really think
Early on, the honest answer is we shipped things too soon. The clearest example was our lesson scheduler, we released it a touch early, customers tried it, hit failures, and complained to us. Worse than the complaints was what it trained them to do, skip self-service and come straight to our customer success team for everything. Once a customer learns your product doesn't work but your people do, that behavior is brutally hard to undo. We're still working on scheduler adoption with a few of our oldest customers today, walking them back into the feature one scheduled lesson at a time.
We learned to under-promise and over-deliver the hard way. We scope commitments tighter now, test new features with customers who know they're testing, and tell customers plainly when something isn't ready. When we ship, it works. Early customers forgive missing features. Being surprised is a different story.
“One of our first customers was a mortgage lender, NBKC Bank. They put a newer loan officer named James through daily practice calls with our AI customers before he went back on the phones. Twelve days into onboarding, he was already hitting monthly targets, outperforming reps who'd been there a year. His sales manager put it best, "James has never put up numbers like this. Not in the year-plus he's been here. Something clicked."”
What most people get wrong about AI-powered Virtual Avatars & Digital Assistants
Most people think AI is emptying out the contact center, and that training human agents is a shrinking problem. They have it backwards. AI is taking the routine calls, which means the calls that reach a human are now the hardest ones, the angry customer, the complex claim, the situation with real money at stake. The easy stuff that used to let a new agent build confidence is gone. They're thrown into the deep end on day one. The human job got harder, not smaller, and most companies are still training agents like it's 2015, with scripts and quizzes, then letting live customers be the practice.
The second thing people miss, companies are deploying AI agents far faster than they can test them. Vendors report their own performance numbers, which is like a car company grading its own crash tests. Nobody would accept that anywhere else. The market doesn't have an AI agent problem. It has a measurement problem.
What's next for TrueCX
What's next is AI Agent Validation, and we're genuinely excited about it. Our beta launches this month, with general availability shortly after. We think this is one of the fastest-growing needs in the market. Everyone's been joking about "AI talking to AI" for years now, and that's exactly what this is.
We built our Intelligent Virtual Customers to train human agents. Turns out the same AI customers that can pressure-test a person can pressure-test an AI agent just as well. Companies are deploying conversational AI faster than anyone can independently check whether it works. We close that gap.
Lonnie's background
This industry is the only one I've worked in. I spent over 20 years in contact center technology and operations before founding TrueCX. I started in contact center analytics at Sprint, then spent nearly a decade at NICE Systems across product management, sales leadership, and running global customer success. After that, I was CRO at Balto, a real-time call guidance company, which put me back in front of contact center leaders every week. So, no, not starting from scratch.
By the time I founded TrueCX in 2024, I'd sat on every side of this market, building the products, selling them, and helping customers get value from them. I was building the thing I'd watched my customers need for two decades.
Biggest lesson building TrueCX
The biggest mistake was underestimating how fast AI would move and how quickly it would reshape the market we were building in. We'd get traction on something, and by the time it was working, the ground had already shifted, and it was time to pivot again. That's a hard thing to live through as a founder. But I'm glad it happened. It taught the whole company a lesson about change that we'll carry with us because AI isn't done reshaping this world. The teams that survive won't be the ones who guessed right. They'll be the ones who can move when they're wrong.
I'd lean into AI from day one, and I don't just mean using it. I mean rethinking what a startup is.
The old startup mantra was move fast and stay scrappy. You couldn't out-resource the big companies or match their polish, so you competed on speed and hustle. That trade-off is gone. AI took it away. Now you have to be all three at once, fast, not resource-constrained, and shipping the same quality a hundred-person team ships. A small team with the right AI tools can do work that used to take a department.
If I started over, I'd build that mindset into the company on day one instead of discovering it along the way. The bar moved. Scrappy isn't enough anymore.
TrueCX at a glance
Website
MRR
The team has requested to keep their revenue hidden
Founded
2024
Employees
11–50
Country
United States
Target market (B2B/B2C)
B2B
Growth model (Product/Sales)
Product led
Uses AI
Yes
Tech stack
Astro,
Cloudflare,
Google Fonts,
Google Tag Manager,
Gravatar,
HTML5,
Hubspot,
Microsoft Clarity,
MonsterInsights,
oEmbed,
reCAPTCHA,
Site Kit by Google,
Tailwind,
Wix Content Delivery Network (CDN),
WordPress,
WordPress Block Editor (Gutenberg),
WordPress Emoji Support,
WordPress REST API,
WordPress XML-RPC,
Yeost SEO Plugin
TrueCX SEO metrics
Domain rating
26
Rating growth (per month)
1.00