Eric Vaughan, CEO of IgniteTech, Khoros, and GFI Software, walked us through the AI-first transformation that replaced 80% of his global team, and the numbers that came out the other side.
Most companies talk about AI adoption like it’s a software rollout. Buy some licenses, run a training session, check the box. Eric Vaughan did something a lot more uncomfortable. In early 2023, with a profitable 30-year-old enterprise software business humming along, he decided to blow it up and start over.
Eric joined us for our 175th episode of Deconstructing Data, and honestly, it was one of the most direct conversations about AI we’ve had on the show. No hedging. No “it depends.” Just a CEO who made a massive bet and is willing to tell you exactly how it went.
What does an AI-first transformation actually look like?
Eric runs IgniteTech, GFI Software, and Khoros (which IgniteTech acquired about three months before our conversation). The company is 100% remote, very global, and manages more than 40 enterprise products. By most measures, it was working.
But when generative AI hit in late 2022, Eric saw something most executives were still treating as a curiosity.
“This new round of AI, really generative AI… represented an existential threat to our company and I would say to everyone’s company, and I really do mean that and I don’t mean it as hyperbole.”
— Eric Vaughan, CEO, IgniteTech
So IgniteTech spent all of 2023 giving its existing team what Eric called “a gift.” Tuition reimbursement. Outside experts like Ethan Mollick from Wharton running workshops. Cash prizes for the best ideas. And then the big one: AI Mondays. For an entire quarter, nobody at IgniteTech was allowed to work on anything other than learning and practicing AI on Mondays, unless there was a live customer fire.
That’s 20% of payroll for a full quarter. That’s a big deal.
Why do employees resist AI adoption at work?
Here’s the thing. The gift didn’t land the way Eric expected. Some people refused outright. Some tried it once, got a mediocre result, and gave up. And some actively worked against the effort.
“It was very much a culture change. It was not just a technical change by any means.”
— Eric Vaughan, CEO, IgniteTech
His favorite example is one I think every marketer will recognize. Someone told him they asked AI to write an article and it didn’t sound anything like them. His response:
“How was it supposed to know you? Did you give it like your last 25 articles that you wrote? How did you give it knowledge so it could sound like [you]?”
— Eric Vaughan, CEO, IgniteTech
I laughed at that one because it’s so true. Garbage in, garbage out still applies.
And here’s where it gets interesting. The group Eric expected to lead the charge was the one that dragged its feet the most.
“Surprisingly, the technical teams were the slowest adopters.”
— Eric Vaughan, CEO, IgniteTech
The pushback from engineers sounded like “it doesn’t know this product like I do.” Which, fair, but it also misses the point. By the end of it, IgniteTech had transitioned roughly 80% of its global team across every function. Then in 2024 they went on a hiring spree, and every new role carried the same prefix: AI Innovation Specialist. Marketing, finance, engineering, all of it. Eric was clear that they weren’t looking for AI experts. They were looking for belief.
What results can a company expect from going AI-first?
This is the part I pushed Eric on, because I think a lot of executives listening need to see numbers before they’ll consider anything this drastic.
He started with support. About 70% of IgniteTech’s level-one support tickets are now resolved in the first session with AI, and customer satisfaction is at an all-time high. His reasoning is simple:
“At the end of the day, most customers are looking for an answer. They’re not looking for a conversation.”
— Eric Vaughan, CEO, IgniteTech
But Eric thinks efficiency stats are the boring part of the story. The real headline is what IgniteTech built. The company closed 2024 at just shy of $3 million in revenue per employee and 75% EBITDA. In that same year, it launched two brand new enterprise products, Eloquens AI (an AI email persona that handles inbound communication) and MyPersonas (AI digital clones), both written by a team that didn’t even work there until the second quarter of 2024.
More recently, after the Khoros acquisition, the team spotted a gap for customers and shipped a working product to fill it in four business days. That’s the kind of statement you could not make four years ago.
His warning to everyone else was blunt:
“If you think you’re behind right now, that’s great. That’s what you should think. If you don’t think you’re behind and you’re waiting to see what happens, I think you’re doomed.”
— Eric Vaughan, CEO, IgniteTech
How can AI speed up mergers and acquisitions?
The Khoros deal is a great case study. Khoros had 650 people when IgniteTech arrived, and Eric’s policy is to meet with every single person one-on-one within about two weeks.
So they built an AI interview tool first. Every employee went through a short AI interview that produced a dossier: role, tenure, reporting lines, who they work with. By the time the human one-on-ones happened, the basic questions were already answered and the conversation could go deeper.
They also deployed Eloquens to answer questions from the outside world about roadmaps and intentions, and to help handle the HR questions that come with any integration. Eric was careful to point out that humans reviewed every one of those emails. They just didn’t have to write each one from scratch. And unlike a tired HR team answering “what happens to my 401k?” for the 200th time, the AI never got short with anyone. It also speaks 160 languages, which matters when your workforce is global.
The result: full operational, engineering, and support responsibility for Khoros within 90 days. Eric’s message to private equity was pretty clear. If you aren’t doing this in your integration model, someone who is will outpace you.
Why does human in the loop still matter with AI agents?
For all his enthusiasm, Eric is not a “let the agents run wild” guy. He actually worries about how the word “agentic” gets thrown around, because it implies you can just hand work off and walk away.
His approach mirrors software QA. You don’t let a coder test their own code, so you don’t let one model grade its own homework either. His team runs outputs from one LLM through another (ChatGPT, Gemini, Claude, Llama) to compare and contrast. He also introduced a term I hadn’t heard before: grounding, meaning the AI can actually confirm that its answer is correct. When Eloquens or MyPersonas can’t reach that confidence, they loop in a human, either by CCing a person on the email or by asking the clone’s owner directly.
On building products, his advice was just as practical:
“Don’t marry yourself to any platform. Don’t marry yourself to any technology. Make it very modular.”
— Eric Vaughan, CEO, IgniteTech
That lines up with what we do at BDEX. I’m a big fan of OpenRouter for exactly this reason. You can swap models on the fly. Eric’s stack includes AWS, OpenAI, Claude, Llama, Mistral, and Google (which his team finds especially strong for code), plus Cursor, which keeps winning their side-by-side comparisons against Claude Code. His other tip: buy the pro versions of everything. Handicapping your team with free tiers to save a few bucks makes no sense.
The bigger picture
I spend more of my day inside AI tools now than anything else, so a lot of what Eric said hit close to home. But what stuck with me most wasn’t the 80% number that grabs headlines. It was his point that AI’s real value isn’t doing the same work with fewer people. It’s doing things you couldn’t do before at all. Shipping software in days. Integrating a nine-figure acquisition in a quarter. Building brand new products with a brand new team in under a year.
You don’t have to agree with every decision Eric made to take the lesson seriously. The companies that treat AI as a culture change, not a tool purchase, are going to move a lot faster than everyone else. And your competitors are going to show up with that speed whether you’re ready or not.
Want to connect with Eric? Follow him as The Gen AI CEO on X and LinkedIn, and check out what IgniteTech is building at ignitetech.ai, khoros.com, and gfi.com.
Want better data powering your AI? AI is only as good as the data you feed it. BDEX delivers the most accurate identity data in the industry. Visit bdex.com and click “Talk to an Expert.”
Watch the full episode: AI-Driven Growth: Restructuring, Strategy, and Scale on YouTube
This article was adapted from an episode of Deconstructing Data, BDEX’s weekly podcast on data-driven marketing. Tune in live every Thursday at 4:15 PM Eastern on LinkedIn.
About BDEX: For companies that need clean identity data to power their products, BDEX offers unmatched quality and execution. Visit bdex.com and click “Talk to an Expert” to get started.
