Sep 15, 2025 · 51m · mixergy
#2280 Read.ai is adding 50k users per day
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Read.ai founder David Shim details how the company achieved viral hypergrowth in a crowded AI productivity landscape by leveraging proprietary multimodal sentiment analysis, cross-platform integrations, and objective post-meeting coaching.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Andrew holds 25.1% of the talking time here. How this is scored →
Andrew as informed peer 4.3 Guest teaching 4.9 Guest disagreement 2.0 Andrew pushing back 3.1
speaking balance: gold is Andrew, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 7:38 Shim dismisses role-specific SaaS tools
Shim bluntly rejects Warner's premise that software should segment into vertical niche tools, asserting that one horizontal AI platform will rule them all.
Hardest push from Andrew ▶ 8:06 Warner insists on vertical departmental workflowsWarner refuses Shim's horizontal platform argument, laying out distinct workflows across his own team's success, sales, and dev roles to challenge Shim's thesis.
Biggest teaching moment ▶ 28:24 Shim breaks down dead reckoning and signal stackingShim educates Warner on how location signal processing models from Placed map directly to decoding conversation sentiment from noisy video and pacing cues.
Andrew holds their own ▶ 35:12 Warner critiques Zoom AI UX and analyzes lock-inWarner demonstrates sharp industry analysis, critiquing Zoom's native AI summaries as tasteless while probing the structural dynamics of cross-platform lock-in.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Andrew as informed peer | Guest teaching | Guest disagreement | Andrew pushing back | Why |
|---|---|---|---|---|---|---|
| Differentiating Through Sentiment and Engagement Tracking | 5 | 4 | 3 | 6 | Warner aggressively questions whether users actually care about sentiment and engagement tracking over basic transcription. Shim pushes back on Warner's premise by arguing that chronological notes are ineffective without real-time emotional metadata. | |
| Horizontal Platform Versus Vertical SaaS Niches | 6 | 6 | 4 | 7 | Warner presses hard on why AI note-taking hasn't fractured into vertical SaaS tools for specific roles like sales, HR, and dev. Shim reframes the question, arguing that legacy SaaS thinking fails in AI because horizontal platforms break data silos. | |
| Founding Read.ai: From Foursquare to Cabo Zoom Fatigue | 3 | 3 | 1 | 1 | Warner asks about Shim's background from Foursquare to founding Read.ai during the pandemic in Cabo. Shim recounts spotting ESPN reflections in video call participants' glasses as the catalyst for measuring engagement. | |
| Unexpected Use Cases: Healthcare, Dementia, and Enterprise Research | 2 | 5 | 1 | 1 | Warner asks who the initial target customer was, expecting sales teams. Shim reveals unexpected emergent use cases including early-onset dementia patients, HIPAA-compliant healthcare field workers, and Mag 7 PM research repositories. | |
| Evolution from Real-Time Distraction to Post-Meeting AI Coaching | 5 | 4 | 2 | 2 | Warner explores whether sentiment feedback was intended in real time or post-meeting. Shim explains how real-time prompts caused cognitive overload and defensiveness, prompting a pivot to post-meeting coaching, which Warner compares to chess.com post-game reviews. | |
| Zoom CEO Validation, Seed Round, and Training Data Challenges | 4 | 5 | 3 | 4 | Warner brings up Shim pitching Zoom CEO Eric Yuan and asks why they did not do customer validation before building. Shim corrects the narrative regarding Yuan and explains that ground-truth training data had to be built from scratch with actors. | |
| Signal Processing Roots: Farecast, Placed, and Dead Reckoning | 3 | 6 | 1 | 1 | Warner inquires into Shim's background in AI. Shim delivers a detailed breakdown of signal processing roots at Farecast and Placed, explaining how dead reckoning of weak and strong signals informs Read.ai's multimodal models. | |
| Building the System of Record and Sales Copilot Workflows | 5 | 5 | 2 | 3 | Warner asks if Read is building an email client like Superhuman. Shim details Read's Sales Copilot and cross-platform integrations that draft follow-up emails, surprising Warner who enthusiastically validates the pain of manual CRM updates. | |
| Platform Risk, Multi-Tool Reality, and Global Virality | 6 | 6 | 2 | 4 | Warner raises platform risk from Zoom, Google, and Microsoft bundling native AI. Shim counters with data showing Microsoft Copilot launches actually drove a 15x increase in Read's Teams usage due to multi-platform user realities and organic virality. | |
| Startup Advice: Shipping Fast, Hallucinations, and Future AI Concepts | 4 | 5 | 1 | 2 | Warner asks for startup advice based on Shim's angel investing. Shim emphasizes shipping fast despite hallucinations, predicting high-value opportunities in AI content authenticity certificates and asynchronous out-of-office AI agents. |
Statements from this episode (18)
Shim: Read.ai adds 50,000 accounts daily as fastest-growing AI note taker
“Yeah, so it's in the millions, and then on a daily basis, we get about a 50,000 new accounts created every single day, so that's a run rate of a million plus on a monthly basis, twelve million annually, so we are the fastest growing meeting note taker in the w…”
Shim: Basic transcription and summarization are commoditized table stakes
“It's very easy to build a basic model. So transcription, you can do open source. You can buy some stuff. You can drop it into chat GPT or Anthropic and get a summary against it. And those are good. And you'll see there's this kind of table stakes of you should…”
Shim: Read.ai tracks head orientation for engagement without facial recognition
“We look at head orientation, so we don't look at your face, we don't do any facial recognition, but we go and we say, if you're looking at the camera straight ahead, and you're talking, your camera's there, and then if you look over this way, and you go to a f…”
Shim: Chronological meeting summaries are ineffective
“So when you get notes that summarize it in chronological order, those aren't effective notes.”
Shim: Read.ai hits 80% 30-day retention for active meeting users
“Most recently in this last month, if you use our product in a meeting, and you get the reports we see that the retention is 80% after 30 days”
Shim: Horizontal AI note takers will win over vertical SaaS tools
“There's one meeting note taker to rule them all in the sense of... A hundred percent. Cause it's too niche. You're not going to go in and say, oh, I've got a sales call. I'm going to invite this one. I've got a call with my doctor. I'm going to do this one. I'…”
Shim: Placed was profitable and growing 50%+ when Snap spun it out
“2019, Snapchat spun us out, not because we were doing poorly, but because we were actually profitable, growing at, you know, 50% plus clip rate.”
Shim: 'Magnificent Seven' companies use thousands of paid Read.ai licenses
“We've got mag sevens using us with thousands of licenses that are paid for where they're aggregating 200 product managers, five customer calls every week per product manager, and they've built this silo or this kind of storage of intelligence of customer inter…”
Shim: Real-time sentiment metrics cause cognitive overload for junior sellers
“The second thing was after the fact, taking those coaching metrics, because a lot of time it is going to be too much cognitive overload, especially for a more junior seller where they're not used to taking all these inputs because you've got a PowerPoint prese…”
Shim: Sales reps accept AI coaching feedback with far less defensiveness
“And they're not defensive because it's the AI telling you. It's not your manager telling you where you're like, okay, is this going to bet my job? This is the AI saying like, make a left turn or right turn. And you're like, okay, totally makes sense. I'm not e…”
Shim: Frequent interruptions after five minutes indicate poor conversational synchrony
“You and I might interrupt each other because we're still getting a cadence. We call it synchrony, but after five minutes, if you're still interrupting each other, that's almost intentional. That's, it's not a good conversation. You both aren't paying attention…”
Shim: Read.ai has connected millions of Gmail and Outlook accounts
“In tiny now for Gmail, as well as for outlook, you can actually connect your accounts and we have millions of accounts that people have actually connected.”
Shim: Read.ai Sales Copilot managed $50M in deal changes in 3 months
“So we've only had that product out for three months. Over fifty million dollars in deals have gone through that where we've initiated a change in status.”
Shim: Vast majority of Read.ai users use multiple conferencing platforms weekly
“So for our users on a weekly basis, we see that they use more than one platform for the vast majority of our users.”
Shim: Microsoft Copilot launch drove 15x surge in Teams usage on Read.ai
“We've seen more than a 15 X increase in Microsoft team meetings that we've measured since Microsoft Copilot launched for teams meetings.”
Shim: 1% to 2% of Colombia's entire population uses Read.ai daily
“We have about one to two percent of Columbia's population using read on a daily basis.”