Real-time speech-to-speech accent harmonisation — a $20M check written before there was a product or a dollar of revenue, then $22M+ of revenue in the 18 months after launch
0 to $20-23M in 18 months · every new customer adds roughly $500K in revenue · $30M round at $150M post while pre-product and pre-revenue, of which Insight wrote $20M · earlier $5M round after two $50K angel checks · roughly 200 people in India · valuation described as more than $500M
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
In 2021, Sharath Narayana was running a quarterly ritual of visiting every enterprise customer at Observe.AI, the contact-centre AI company he had cofounded. In one of those visits a CIO named Prashant told him he had seen something magical the previous week: some very early Stanford founders claimed they could put software on an agent's desktop that harmonised their accent in real time — "but again, this is probably still one or two years away." He meant to reach out that same day and forgot. Two weeks later he was in the Stanford speech lab and heard a professor describing the same concept. He asked whether they meant Sanas. They did. He met Andreas and Max — who were 22, 21 and 20 with Shawn when he first met them — and invested in their first round. What pulled him in was technical, and specific to a problem he already had. At Observe.AI they were fighting perceivable latency in real-time speech. Sanas had found a way around it by not waiting for speech-to-text and then text-to-speech at all: instead, phonetic mapping on parallel pairs in real time, so you never have to wait for a whole word to be spoken because a phoneme is the atomic unit of sound. His first advice was not commercial: "first patent this, take my money, go and file for a patent." He and Steve at DN Capital wrote $50,000 checks each, then helped construct a $5M round with General Catalyst, Quiet Capital and Human Capital. Max and Shawn then asked him to come on board full time. He joined in January 2022, leaving a company that had gone from zero to about $15 million. Within 60 days, Insight Partners wrote a $20 million check into a company that was still pre-product and pre-revenue; the total round was $30 million at $150 million post. Then two more years of building before market entry, and $22M+ of revenue in the 18 months since.
6 more sections — the full story, the playbook, the risks, and the numbers. Free account unlocks everything.
Create free accountEmail code only. No password.The quick-reference card — acquisition channels, replication playbook, and risk map — unlocks with a free account.
Create free accountEmail code only. No password.Paste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea: a full build spec and an SEO growth plan. Copy, paste, ship.
Go Pro — $10/moCancel anytime.Every claim traces back to a source video.
Source video links unlock with a free account. Full transcripts are Pro.
Create free accountEmail code only. No password.Data credibility: Founder interview on a venture fund's own podcast with a disclosed two-way commercial relationship: the host is managing partner of Neon Fund and the interviewee states he is an investor in Neon. No dashboard, filing or third-party data appears; every figure is self-reported and unaudited. Self-reported figures: $22M+ in revenue in the 18 months since going to market (stated at one point) and 'we went from 0 to 20, 23 million in 18 months' (stated at another) — we have recorded both rather than merging them, since they may be different measurements or dates; roughly $500K of revenue added per new customer; a $30M round at $150M post while pre-product and pre-revenue, of which Insight Partners wrote $20M, closed as a term sheet in March 2022 and announced in June; a prior $5M round with General Catalyst, Quiet Capital and Human Capital, preceded by two $50K angel checks from Narayana and Steve at DN Capital; roughly 200 people in India; and a valuation the host describes as more than $500M for each of the interviewee's two companies. Technical claims — sub-200-millisecond speech-to-speech via phonetic mapping on parallel pairs — are as stated by the founder and unverified here. The founders' ages when first met (Max 22, Andreas 21, Shawn 20) and the timeline (introduction 2021, Narayana full time January 2022, Insight within 60 days, market entry two years later) are all his account. The near-failure detail is his own disclosure: eight to ten months into the $30M they discussed with Insight giving it another 90 days before returning the capital. IMPORTANT NAMING CAVEAT: the auto-captions never render the company name reliably, producing SAS, sis, sers, sonus and sanus; 'sanus' and 'sanas' both appear and we have used Sanas. Region is the United States: the company was founded by Stanford students, its investors and go-to-market are US-based, and the interviewee describes the product as sold to North American enterprises — while roughly 200 of its people are in India, which is where the software is built. The 20-million-overseas-agent figure and the BPO relocation argument are relayed secondhand from an executive named only as 'Andy from AA' in the captions, and we have not attempted to identify the company.