The interviewer who never gets tired
The first version of Strella was a person pretending to be a robot. Co-founder Priya Krishnan would join a Zoom call with her camera off, switch to a flat synthetic voice, and wait a deliberate five seconds before answering - a hand-built impression of machine latency. The question Lydia Hylton wanted answered was not technical. It was behavioral: would a stranger sit across from something that clunky and actually open up?
They did. People talked, they stayed, they confessed things. The social friction everyone assumed would kill an AI interview simply never showed up. That single observation became a company. Today Strella runs AI moderators that hold adaptive, follow-up-heavy conversations up to 90 minutes long, then synthesizes the transcripts into insights teams can act on the same afternoon. The work that used to eat eight weeks now fits in an afternoon.
Four careers, one bottleneck
Hylton arrived at this problem from an unusually wide approach. She studied computer engineering at Georgia Tech as a Stamps Scholar, then crunched numbers as a data scientist at Facebook. After a Stanford MBA she moved into venture capital at Redpoint, and in 2022 became a partner at Bain Capital Crypto, a $650M Web3 fund where she backed early-stage DeFi and consumer protocols. Each seat - consultant, engineer, operator, investor - kept circling the same complaint.
"I would run surveys, I would run expert interviews until the cows come home."
Quality interviews were slow, stretching into weeks or months. Surveys were fast but riddled with fraud and respondents rushing to the finish. Companies that wanted to stay close to their customers kept hitting the same wall, so they skipped the research entirely. Hylton and Krishnan - friends since high school, entrepreneurs in their heads since then - founded Strella in July 2023 and spent a full year in what they call the idea maze before committing to AI customer research in mid-2024.
Selling to strangers on the internet
What followed was a discipline most early startups skip. Instead of leaning on warm introductions, Hylton and Krishnan recruited twelve design partners entirely through cold LinkedIn outreach. The partners met biweekly, used raw early builds, gave structured feedback, and accepted a hard bargain: a 50% discount for six months in exchange for a deadline to convert to paying or walk. When the deadline came, all twelve stayed. A perfect conversion rate is rare. A perfect conversion rate from cold strangers is the kind of evidence venture investors dream about.
"It's much harder to validate an idea if you need to sell a product to a stranger on the internet."
The numbers backed the instinct. Strella hit $1.6M in annual recurring revenue in its first year of monetization, and its first renewal cohort posted net dollar retention north of 150% - customers not only stayed, they spent more. Since coming out of stealth in October 2024, the company quadrupled its customer base and grew revenue tenfold.
Who's on the line
The customer list reads like a brand-strategy off-site: Amazon, Duolingo, Chobani, Square, Daily Harvest, Apollo GraphQL. Duolingo used Strella to wrap a video concept test in two days that previously took six-plus weeks. The pricing is deliberately unfussy - seat-based rather than per-project - because Hylton wanted to delete the recurring budget-approval conversation that quietly strangles internal research.
The longer bet
In October 2025, Bessemer Venture Partners led Strella's $14M Series A, joined by Decibel Partners, Future Back Ventures by Bain & Company, MVP Ventures and 645 Ventures. The new money goes toward mobile research, a bigger engineering team, and methods that only become possible once an AI is doing the asking. Hylton's thesis runs against the grain of the AI moment: the more decisions machines make autonomously, the more valuable grounded human testimony becomes. Strella's endgame is a fully agentic research loop, where AI gathers what real people think and routes it straight to the systems making the calls.
"We built Strella to make research faster, smarter, and more human - proving that teams no longer have to choose between speed and depth."
There's a smaller story tucked inside the big one. When Hylton joined Bain Capital Crypto in 2022, she wasn't in the fund's launch announcement - she hadn't yet told Redpoint she was leaving, a detail TechCrunch wrapped around "that tweet." It's a tidy preview of the founder she'd become: willing to make the unconventional move, comfortable explaining it later, and unbothered by being early to something the rest of the room hasn't priced in yet. Crypto then, AI interviewers now. The constant isn't the technology. It's the appetite for a bet other people find strange.