Kurt Kratchman runs a company built on an old frustration: most workplace training does not stick. People sit through videos, click through slides, pass a quiz, and forget it by the following week. At Virti, the New York AI company he leads, the pitch is simpler than it sounds. Instead of watching a scenario, you step into one and talk your way through it with a virtual human that talks back.
Virti is an AI role-play and interactive video training platform. Its software builds lifelike digital characters - virtual humans with faces, voices and moods - that learners can practice with using natural speech. A sales rep can rehearse a tough negotiation. A manager can work through a difficult review. A clinician can walk into a simulated patient encounter and make the call, get it wrong, and try again without anyone watching over their shoulder. The system listens, adapts, and gives feedback. As CEO, Kratchman's job is to turn that idea into something enterprises trust and buy.
The bet underneath the product is that the hardest skills to teach are the human ones - communication, judgment, empathy under pressure - and that those skills only improve with repetition. Reading about a hard conversation does little. Having one, over and over, in a safe environment, does a lot. Virti's technology stack, blending large language models, computer-generated avatars, 360-degree video and learning analytics, exists to make that repetition possible at scale.
The learner's journey
Ask Kratchman how the product actually works on someone and he reaches for a phrase he uses often: the learner's journey. In his telling, it moves in three steps. First comes situational awareness - dropping a person into a scene so they register what is happening. Then context, so they understand the stakes and the people in front of them. Then action, where they respond and live with the result.
When you get into the immersiveness of the video and you get into the diagnostic tools of the avatars, and you bring them together, that's where the magic happens.
It is a framework that reveals how he thinks. Kratchman is an operator before he is an evangelist. He does not talk about immersive learning as a gadget. He talks about it as a path from confused to capable, and about the analytics that prove someone actually got better. That instinct - measure the thing, then improve it - traces straight back to where he came from.
A career that moved sideways to move up
Kratchman did not start in software. His first job was in traffic engineering in New York City, working out how vehicles move through a grid. From there he crossed into data visualization, rich media, content management and digital distribution - the plumbing of the early internet economy. Along the way he was part of two agency exits to the advertising giant WPP, with Blast Radius and Schematic.
His defining chapter came at Grapeshot, a contextual advertising company where he served as Chief Operating Officer and ran global operations across offices in New York, Cambridge, London and Singapore. When Oracle acquired Grapeshot, Kratchman went with it, becoming a Group Vice President responsible for a portfolio of martech and adtech software-as-a-service products inside one of the largest technology companies in the world.
That is a particular kind of resume. It is not the founder-in-a-garage story. It is the story of someone who knows how to take a promising technology and build the operational machinery - sales, structure, global teams - that turns it into a business. After Oracle, he moved into investing, becoming a founding partner at Makude Investment Partners and working as a venture partner at Winnick & Company, backing technology ventures alongside colleagues from his Grapeshot and Oracle days.
The learner's journey starts with situational awareness. They get into the context of the situation, they see what's going on.
Why Virti, and why now
Kratchman joined Virti around its Series A raise, first as Chief Revenue Officer, then stepping up to CEO. It is worth being precise about the company's origins: Virti was founded by Dr. Alexander Young, who brought the medical-education vision. Kratchman is the operator brought in to scale it. That division of labor is common in growth-stage companies and it plays to exactly what he has spent his career doing.
The timing was not incidental. The pandemic broke a lot of how organizations trained people. Classrooms closed. Shadowing stopped. In healthcare especially, where practice on real situations is not optional, the gap was urgent. Virti's answer - remote, immersive, repeatable practice through VR headsets, 360 video and avatars - suddenly looked less like a novelty and more like infrastructure. The pandemic eased. The underlying need for scalable, hands-on training did not.
Then came the generative-AI wave, and it reshaped what Virti could build. The company leaned into AI role-play, using large language models to make its virtual humans converse more naturally, and pushed to make those characters more realistic and more diverse - varying body types, ages, ethnicities and clothing, so a learner practicing a patient conversation or a customer complaint sees someone who could plausibly be real. For a company whose whole premise is believability, realism is not cosmetic. It is the product.
The humanities major who runs a tech company
There is a detail in Kratchman's background that resists the usual tech-executive template. He studied philosophy and Jewish studies at Tulane University, spent time at The Hebrew University of Jerusalem, and later trained in new media at the Vancouver Film School. It is a humanities-heavy path into enterprise software, and it fits the work more than it might first appear. Virti is, at bottom, a company about stories and how people absorb them. A leader who has thought about narrative, meaning and media is not an odd fit for a business built on scenario-based learning.
He also founded and led BLACKBOX, an aviation-related venture, another reminder that his career has moved across industries rather than up a single ladder. The through-line is not a sector. It is a question he keeps returning to in different forms: how does information actually move through people, and how do you make it land.
The best training does not feel like training. It feels like practice - the kind you can fail at, safely, until you get it right.
The case, and the skeptics
The obvious question hangs over the whole category: can you really learn a human skill from a machine. Kratchman's answer is not that AI replaces the human moment. It is that AI gives you a place to rehearse before the human moment arrives - a flight simulator for conversations. A nurse who has practiced delivering hard news a dozen times with a virtual patient walks into the real room differently. A new manager who has run the awkward performance review in software is less likely to freeze in the actual one.
Whether enterprises buy that at scale is the test in front of him. Virti competes in a crowded learning-and-development market against everything from legacy training vendors to a wave of AI-native startups. The company is small - a lean team, a modest raise by the standards of the AI gold rush - and it is selling a behavior change, not just a tool. Kratchman's operator instincts, honed at Grapeshot and Oracle, are the reason investors and colleagues have followed him here. The pitch is that he has done the hard part before: taking something clever and making it a company.
He is candid that the category has to earn its credibility. Immersive learning has a history of overpromising - shiny VR demos that impressed a boardroom and then gathered dust because no one measured whether they changed behavior. Kratchman's counter is data. Virti tracks how learners perform inside scenarios and folds that into analytics managers can actually use, so the argument for the platform is not that it feels futuristic but that it produces measurably better performance. In a budget conversation, that distinction is everything.
There is also a practical argument he keeps close: content has to be easy to make. A training tool nobody can author is a training tool nobody uses. Virti has pushed toward no-code scenario creation and generative content so that a learning-and-development team, not a studio, can build and update simulations. It is a quiet but decisive bet - that the winners in AI training will be the platforms non-technical teams can run themselves.
For now, he is doing what he has always done - taking a technology with obvious promise and building the unglamorous machinery around it. The subject has changed, from advertising data to virtual humans. The job has not. Find the thing that works, prove it works, and make it scale.