Conviction Founder Sarah Guo: Video Conferences Fail to Accurately Read Founders' Signals
Sarah Guo, founder of Conviction, commented on Rogo founder Gabe Stengel, stating that she is not good at reading signals from video conferences: she first met Gabe online, and later realized his potential when they met in a social setting in New York, calling it a lesson learned repeatedly.
In her statement, she did not announce any additional or missed investment in Rogo, but attributed judgment errors to the medium: remote meetings compress the presence of founders. In the same discussion thread, Gabe recounted that around 40 institutions had met, gone through investment committees, dinners, and even weekends before ultimately rejecting him, with Keith Rabois leading a $18.5 million round from Khosla Ventures.
Sarah Guo left Greylock in 2022 to establish Conviction, an early-stage fund focused on artificial intelligence, with the first fund at approximately $101 million and the second fund at about $230 million, bringing in former Sequoia partner Mike Vernal. At 28, she became Greylock's youngest general partner, participating in Figma's Series A and engaging with what later became TikTok, Musical.ly. Her public portfolio includes legal AI Harvey, reasoning platform Baseten, doctor-facing OpenEvidence, customer service Sierra, programming Cognition, and open-source model company Mistral.
Harvey was the company Rabois casually compared Rogo to. Guo is one of Harvey's early supporters, which operates within law firm workflows; Rogo operates within investment banking, private equity, and asset management workflows. Both were founded by former junior professionals who turned their "hard work" into vertical software. Rogo has since raised over $300 million, with a post-Series D valuation of about $2 billion, and publicly reported user numbers exceeding 35,000, covering over 250 institutions.
Guo also hosts the podcast No Priors, and her investment style is summarized by her as being few and concentrated: over three years, she has invested in about 27 companies, serving on the board of only six. She views early investments as a long-term suspended skepticism about a person and an idea, rather than completing signal confirmation in video meetings.
In market mechanisms, this represents cognitive correction rather than a new round of pricing. Buyers still need to convert founders from pixels back into collaborative entities, while sellers must navigate through the Zoom filtering funnel. The event-driven aspect comes from the mutual verification within the circle after the same interview was shared. The beneficiaries are those who can move due diligence to New York dinner tables, while the pressured parties are institutions that consider a one-hour video sufficient for decision-making; Rogo's equity has already been redistributed among Khosla, Thrive, Sequoia, Kleiner, and strategic capital from banks.
Source: Public Information
ABAB AI Insight
Sarah Guo's historical actions have always been about "seeing people clearly before consensus forms." A month before the ChatGPT explosion, she bet her entire career on Conviction, with her portfolio including Harvey, Sierra, Cognition, and Mistral, all of which later became billion-dollar narratives. Her ten years at Greylock provided her with the filtering muscle for software and infrastructure, and leaving the large platform was to narrow her decision radius to a few people she is willing to suspend skepticism about long-term. Her comment on Gabe acknowledges that the same muscle fails in video channels.
The capital path did not move new dollars in this post; it moved the medium of due diligence. Conviction's funds are still primarily stacked in verticals like legal, reasoning, medical, and customer service that she has already validated; the financial version of Harvey was ultimately priced by Khosla, then picked up by Sequoia and Kleiner at a higher valuation. Guo treats in-person meetings as a calibrator, indicating that her fund still views "people" as an incompressible asset, rather than treating pitch links as scalable pipelines.
The analogy is with Harvey, which she has invested in: Winston Weinberg is also a former junior lawyer who productized document work, and early on also needed investors to complete belief leaps in the office rather than in front of a screen. In terms of industry positioning, vertical AI has shifted from competing on demos to competing on on-site deployment; projects that can pass video screening are increasing, while those that can persuade investors at the dinner table remain scarce.
The structural change represents information loss in the transfer of pricing power. Remote meetings have lowered the cost of meeting, but also the cost of rejection, leading to a pipeline where A rounds see "discussions continuing into the weekend before rejection." Whoever can restore founders from compressed video to collaborative density will act first among the same batch of rejected projects. The mechanism is rigid: signal bandwidth determines who pays first, not the business model canvas.
ABAB News · Cognitive Laws
Video reduces the cost of meeting but also lowers the resolution of seeing people.
Early checks buy density in the same room, not pitch links.
Investors misjudging the medium is often more costly than misjudging the track.