Overview
Torch turned a human coaching practice into an employer-funded development platform. Founded by Cameron Yarbrough and Keegan Walden in 2017, it joined YC’s Winter 2018 batch. It added assessment software, combined with Everwise, and later acquired AI learning company Praxis Labs.[1][5][9]
The later outcome is an acquisition with a continuing advertised product. Former CEO Heather Conklin says Torch was acquired by Xenon Partners and that she stepped down in May. Jason Gilmore separately announced the acquisition. The observed statements leave the closing year, transaction terms and investor returns undisclosed.[10][11]
Torch’s trajectory tests a service-to-software strategy. Assessments and administration can make coaching easier to buy at scale. They cannot, by themselves, prove that coaching caused a business outcome. Its current human-plus-AI offering also changes the rebuild premise: routine conversation practice is already part of the incumbent product.
Image 1 / 1
Founding Story
The founders met while studying counseling psychology at the California Institute of Integral Studies. Walden’s account describes training with Yarbrough, then pursuing a doctorate while Yarbrough developed a clinical practice and the Well Clinic. A July 2017 conversation brought them back together around professional development.[2]
Their first customers were founders and executives they coached directly. Walden says this practice exposed a recurring information problem: a coach initially hears the leader’s own account, while leadership is experienced by colleagues. Torch’s early 360 tool asked the leader, manager, peers and direct reports about the same behaviors. Differences between those accounts gave the coaching work a more concrete starting point.[2]
Yarbrough described email matching, Zoom sessions and monthly invoices before the platform.[4] The useful sequence was a delivered service followed by software for its recurring coordination problems. That account supports the development sequence; it does not provide audited churn or cohort economics.
Timeline
- 2017–2018: Torch is founded and joins YC Winter 2018.[1]
- April 2019: Torch raised $10 million and reported 100 clients.[3]
- March 16, 2020: Torch announces its combination with Everwise under the Torch brand.[5]
- February 25, 2021: A $25 million Series B brings reported total financing to nearly $40 million.[6]
- May 2022: Torch announces a $40 million Series C, led by 137 Ventures.[7]
- August 1, 2024: Heather Conklin becomes CEO; Yarbrough becomes executive chair in an announced succession plan.[8]
- July 10, 2025: Torch announces its acquisition of Praxis Labs, adding AI practice and skills assessment capabilities.[9]
- Later acquisition disclosure: Conklin reports the Xenon Partners acquisition and her departure. The exact closing date remains unverified.[10][11]
- Current offering: Torch advertises human coaching, Spark AI practice and program reporting. YC still labels it inactive.[12][1]
What They Built
Torch combined coach matching, assessments, goals and program administration. The 360 assessment addressed a limitation of private coaching: the participant’s self-description is only one perspective. Program tools let the employer organize development without having to coordinate every relationship manually.[2]
Everwise expanded that operating layer. The 2020 announcement described mentoring, peer support and facilitated learning, with a global community of more than 6,000 mentors. It named customers including Gilead Sciences, Best Buy, T-Mobile, Zendesk and Genentech. This was a historical mentor network, not a current count of Torch coaches.[5]
Today’s product page shows participants comparing three proposed coaches, including a recommended match. It also describes organizational insights and coaching themes for program managers. Some dashboard and financial examples are explicitly demonstrations, so their displayed results must not be treated as customer evidence.[12]
Praxis added a second delivery format: repeated AI practice between, or without, human coaching sessions. Torch now distinguishes unscored Spark rehearsal from scored skills practice with scenario rubrics. Its current practice page says transcripts are private to participants, audio is not stored, and administrators can request aggregate engagement and skills data.[13]
Image 1 / 1
Market Position
Torch sells development programs to employers. Its 2021 funding release named enterprise customers and described consolidating learning tools for learning-and-development teams. The buyer’s job widened from arranging coaching for a few executives to administering several formats across an organization.[6]
Zendesk’s Ignite program makes that breadth concrete. Torch’s current account describes an eight-month experience combining external mentoring, group coaching, learning and a parallel manager pathway. Reported midpoint feedback concerns motivation and satisfaction. The page does not provide respondent counts or prove a causal improvement in business performance.[18]
Competition includes human coaches, managed coaching platforms and AI practice. BetterUp announced its AI offering in January 2025; CoachHub announced its AIMY release in February. Torch’s own Praxis acquisition and Spark product place it within the same hybrid category.[16][17][9] The presence of competing launches establishes alternatives, not demand for another vendor or negligible delivery cost.
A December 2023 Gartner survey found one-third of HR leaders planned to increase learning-and-development spending in 2024. That is dated buyer-intent evidence, not a current coaching-market size or realized purchase total.[15]
Business Model
Torch sells employer-funded coaching and development programs. Its funding announcements show investment in product capabilities and go-to-market operations. The human service still requires coach capacity, matching and quality management; AI practice adds its own evaluation, privacy and operating costs. Whether the bundle produces attractive margins depends on contract terms and delivery costs that the public record does not disclose.
The three named rounds total $75 million. The 2021 release places cumulative financing near $40 million before the Series C. The listed rounds alone do not establish an exact all-time funding total. These are funding measures, not revenue or valuation.[3][6][7]
In 2022, Torch said revenue and employee count had more than doubled year over year, and it named Zynga as a new customer and Zendesk as an expanded relationship. It also reported 38% higher retention among coached employees and 15% higher team engagement compared with an uncoached cohort. Those vendor comparisons do not disclose enough about selection and controls to isolate a coaching effect.[7]
Current organizational-intelligence material describes aggregating 360 patterns and development themes. Such reporting can help procurement assess a program. A before-and-after score or a synthetic revenue illustration still requires a separate test of attribution, costs and missing data before becoming an ROI claim.[14][12]
Post-Mortem
Torch’s observed outcome is more specific than a directory label: a disclosed acquisition by Xenon Partners, alongside current product marketing. The available statements do not provide a purchase price, ownership structure, closing date or return to investors. Conklin’s description of strong headwinds is her account of the period, not a complete explanation of the transaction.[10][11]
The strategic mechanism was expanding the coordination layer around a human relationship. Everwise added formats and administration; Praxis added AI practice and assessment. Each deal could extend what an employer could buy from one vendor. It also required integrating different delivery systems and deciding what information participants, coaches and employers should receive. The announcements establish the changes, while the effect on margin and renewal remains unknown.
One possible pressure is the cost of human delivery as participation grows. A countercase is that employers pay for experienced judgment precisely where generic practice is insufficient. Torch’s continuing human-plus-AI product is consistent with that distinction. Neither explanation establishes why Xenon bought the company or whether the venture-growth strategy succeeded financially.
The evidence also limits the proposed rebuild. Private conversation practice, human coaching and aggregate reporting are existing capabilities. A focused six-week manager cohort could still be useful, but its value must come from a buyer-tested program, delivery quality or purchasing arrangement. Merely adding an AI roleplay cannot establish differentiation.
Key Lessons
- Build around an observed service problem. Torch’s assessments addressed the gap between a leader’s self-image and colleagues’ experience.
- Keep the units behind growth claims. Historical mentors, current coaches, employer accounts, financing and revenue are different measures.
- Separate program reporting from causal proof. Survey movement can inform a buyer without proving retention gains or financial returns.
- Recheck the incumbent before rebuilding. Torch’s own acquisitions and Spark product already cover the generic human-plus-AI practice concept.
Sources
- YC identity
- Walden founder interview
- Torch Series A
- Founder concierge account
- Torch and Everwise announcement
- Torch Series B
- Torch Series C
- Torch CEO succession
- Torch Praxis Labs acquisition
- Heather Conklin acquisition account
- Jason Gilmore acquisition announcement
- Torch current product
- Torch Spark practice and privacy
- Torch organizational intelligence
- Gartner dated buyer survey
- BetterUp AI launch
- CoachHub AIMY launch
- Torch Zendesk Ignite program