The people development platform behind Reddit, Zendesk, and Twitch. Two psychologists, $85M raised, and a quiet bet that leadership is a skill you can actually teach.
Walk into a 2026 enterprise people-ops review and you will find, somewhere on the slide deck, a Torch logo. Often more than once. The company has spent the last eight years quietly turning executive coaching - that vague, expensive thing companies used to buy for their CEO and one or two heirs apparent - into software with seat counts, dashboards, and a renewal cycle. It is not the most romantic story in SaaS. It is, increasingly, one of the more durable ones.
Torch sells what it calls a people development platform. Underneath the label sits a stack of services that used to live in three different invoices: 1:1 executive coaching, manager coaching, group programs, mentoring software, and a small mountain of measurement. The company has roughly 400 employees, a network of vetted coaches working through its platform, more than a hundred enterprise customers, and $71.8 million in 2024 revenue. Those are not unicorn numbers. They are something subtler: the numbers of a company that has found a category that doesn't churn.
The CEO, Cameron Yarbrough, trained as a clinical psychologist. So did his co-founder, Keegan Walden. They met in graduate school, ran group therapy sessions together, and somewhere between Stanford GSB and Y Combinator decided the corporate world was a kind of clinic that didn't know it was sick. The pitch was simple enough to sound suspect: leadership is a learnable behavior, and behavior is measurable, and if you measure it long enough you can attach it to a P&L. Investors eventually agreed. So did Reddit, Zendesk, Twitch, FICO, Tripadvisor, Zynga, Gilead, T-Mobile, and LinkedIn.
The dirty secret of management is that almost no one is trained for it. The salesperson who hits quota gets a team. The engineer with the cleanest commits gets six reports. Then HR sends them a Slack about a leadership workshop in Q3 and considers the matter closed. Predictably, the new managers struggle. Their teams notice. Engagement scores dip. The best people leave. The cycle starts again, with new hires and the same playbook.
Executive coaching, the old fix, was effective and almost cartoonishly inaccessible: $400-800 an hour, scheduled by an assistant, available only to people whose calendars warranted assistants. For everyone else - the actual middle of the org chart, where most of the engagement damage happens - there was a stack of off-the-shelf e-learning modules. The kind people skip while they're on calls.
So you had a market with a clear gap: real coaching, at scale, with proof. Plenty of consultancies had tried. Most became boutique shops with five-figure project fees and no software. Torch's bet was that the missing ingredient was infrastructure - matching, scheduling, progress tracking, 360 feedback, learning content, and aggregate reporting - all wrapped around a global network of vetted human coaches.
Yarbrough's path is the kind of resume that, in hindsight, looks inevitable. Years in clinical psychology. A team of more than 50 psychologists and psychotherapists at his previous venture. A teaching role at Stanford Graduate School of Business where he ran leadership facilitation. Each step nudged him further into the same question: what if the techniques therapists use - real listening, structured reflection, behavior change rooted in evidence - were available to the average mid-level manager?
Walden brought the operating discipline and the science vocabulary. Together they founded Torch Leadership Labs in 2017, walked it through Y Combinator, and started building the kind of company that doesn't usually make the front page of tech blogs: B2B, enterprise-priced, not viral.
It is, of course, slightly funny that the company most associated with "trusted relationships at scale" started as two graduate students practicing group therapy on each other. But the founders never pretended otherwise. The behavioral-science vocabulary is right there on the careers page. The product team uses words like "intervention" and "outcome" without apology. The vibe is less "growth hacking" and more "office-hours-with-your-favorite-professor".
The Torch platform looks, at a glance, like an HR tool. Log in, get matched with a coach, schedule sessions, follow a learning path, leave a 360 assessment, watch a dashboard light up. Underneath, though, it is doing something stranger: tracking a longitudinal behavior-change program for an entire workforce and quietly proving (or failing to prove) ROI back to a customer's chief people officer.
The catalog covers most of the people-development map:
What people actually do with it varies by employer. A scaling tech company might use Torch as their default manager-development track. A regulated enterprise might run a contained executive program for the top 200. A retailer might lean on the group cohorts. The platform's ambition is to be the place every one of those use cases lands.
*Vendor-reported; comparison to uncoached cohort on the Torch platform.
The official mission language - unlock the potential of every leader by combining coaching, mentoring, and learning into a measurable, scalable platform - is the kind of sentence you have heard from a hundred HR vendors. What sets Torch apart, plausibly, is the founders' insistence that the work itself is clinical. The platform is built around a thesis with citations: behavioral change is real, it is measurable, and a trained human listener is still, in 2026, the most reliable intervention we have for it.
That's why Praxis Labs matters. Adding AI immersive practice to the stack is not a generative-AI vanity play. It is the bet that the next decade of leadership development looks like: a human coach plus an AI simulation plus a group cohort plus data. Less of any one ingredient gets you the same workshop everyone has been ignoring for a decade.
Return to that 2026 enterprise people-ops review. The slide with the Torch logo on it. A year ago it might have been the only vendor on the page. Today the deck is busier - competitors have shown up, AI features have crept into every HR tool, the budget conversation is sharper. And yet the Torch line item is still there, with a renewal date next to it. The chief people officer doesn't fight for it. They protect it.
That's the quiet result of building in the unglamorous middle. Not a category-defining IPO. Not a moonshot. A company that bet that humans, helped by software, would keep paying for the chance to get a little better at their jobs. Eight years in, the bet still looks reasonable.
And the founders, the two psychologists who once led group therapy together in grad school? Still arguing about the same things. The vocabulary just spread.