Read AI: Meeting Engagement Scoring, Search Copilot, and Pricing

Read AI, founded by former Foursquare CEO David Shim in 2021, started as a meeting analytics tool built around a proprietary Engagement Score — a metric combining talk time, sentiment, question-asking frequency, and body language read from video when available, to rate how actively each participant contributed to a meeting. By 2026, the company's lead pitch has shifted beyond that original analytics layer toward Search Copilot (also called Ask Read), a unified natural-language search surface across meetings, email, chats, and connected documents in Google Drive, OneDrive, Slack, and Teams, plus a newer email-based "digital twin" feature called Ada.

The company has raised roughly $81 million, including a $50 million Series B in late 2024 at a reported $450 million valuation, and reports adoption across 75% of the Fortune 500 alongside more than 1 million new monthly users added organically without paid marketing. For someone evaluating Read AI, the deciding factor is whether engagement scoring and cross-communication search depth justify a meaningfully higher price than Fathom's free tier, given that Read AI's core sentiment-analysis feature raises a genuine EU AI Act compliance question for European teams, and given specific feature criticism from reviewers who've used it at length.

··········

HOW ENGAGEMENT SCORING AND SEARCH COPILOT ACTUALLY WORK.

A multi-signal engagement metric, unified cross-communication search, and an expanding agentic layer define the mechanism beyond meeting transcription.

The Engagement Score is Read AI's original technical differentiator: rather than just transcribing a meeting, it combines talk-time balance, sentiment analysis, how often each participant asks questions, and — when video is available — body-language signals into a single score meant to indicate how actively someone participated. That's a meaningfully different technical claim than a plain transcript-and-summary tool makes, since it's inferring something about participant behavior and emotional state rather than just recording what was said.

Search Copilot extends the same underlying data collection across a person's broader communication stack: it unifies meetings, email, chats, and connected documents into one natural-language query surface that returns cited answers, with paid plans getting direct in-product access to premium models — OpenAI's GPT-4.1 and Anthropic's Claude 3.7 Sonnet — at no additional charge on top of the subscription. An agentic workflow suite adds recap, catch-up, and scheduling agents aimed at reducing how many meetings someone needs to attend directly, and CRM integration with Salesforce and HubSpot, along with Notion and Slack push, is included starting at the Pro tier rather than gated to the top enterprise plan.

........

Component	Mechanism	Function
Engagement Score	Talk time + sentiment + questions + video body language	Rates participant engagement, not just transcript accuracy
Search Copilot (Ask Read)	Unified NL search across meetings, email, chats, docs	Returns cited answers spanning a person's full communication stack
Model access	GPT-4.1 and Claude 3.7 Sonnet included on paid plans	Adds premium-model reasoning at no extra charge beyond the subscription
Ada	Email-based "digital twin" feature	Extends the product's scope beyond meeting-specific analytics
Agentic workflow suite	Recap, catch-up, and scheduling agents	Aims to reduce the number of meetings a person needs to attend
CRM/Slack integration	Included from the Pro tier, not Enterprise-only	Extends meeting data into existing tools without a premium-tier requirement

........

··········

WHY THE ENGAGEMENT SCORE AND SOME NAMED FEATURES DRAW SPECIFIC CRITICISM.

Emotion and sentiment scoring in a workplace context raises a genuine EU AI Act compliance question, and reviewers name specific feature weaknesses beyond general polish complaints.

The Engagement Score's reliance on sentiment and body-language inference is a real regulatory consideration for European organizations specifically, not just a design preference: at least one detailed 2026 review flagged the sentiment analysis as non-compliant with the EU AI Act, which restricts AI systems that infer emotions in workplace and education settings in most circumstances. A European team evaluating Read AI for its analytics layer specifically should confirm current compliance status directly rather than assume a US-built product's emotion-scoring feature clears EU regulatory requirements by default.

Beyond the regulatory question, reviewers who tested the product at length name specific, non-generic complaints: the "Charisma" coaching feature was described as vague and unhelpful rather than actionable, sentiment readings themselves came across as tone-deaf in some conversations, and removing Read AI's bot from a meeting retroactively was described as frustrating rather than straightforward. None of these are fatal flaws, but they're concrete enough to test directly during a trial rather than assume away based on the product's broader ambition and funding.

··········

HOW READ AI'S PRICING COMPARES TO OTHER MEETING INTELLIGENCE TOOLS.

Read AI positions itself on premium integration depth and governance rather than free-tier generosity, and its pricing reflects that directly.

Read AI's free tier caps at just 5 meeting summaries per month with no CRM sync — a notably thinner starting point than Fathom's unlimited free recordings and summaries, covered separately on this blog. Pro runs $19.75 per month billed monthly (or $15 annual) for unlimited meetings, full engagement scores, and CRM integration; Enterprise runs $29.75 monthly ($22.50 annual) adding video playback and SSO; Enterprise+ runs $39.75 monthly ($30 annual, 10-seat minimum) adding DLP controls, custom retention, and audit logs.

........

Tool	Free tier	Entry paid price
Read AI	5 meeting summaries/month, no CRM sync	Pro $19.75/mo monthly ($15 annual)
Fathom	Unlimited recordings, transcription, AI summaries	Free at core; paid tiers add CRM depth
Fireflies	Limited transcription and AI summaries on free	Reported 15–20% below Read AI's entry pricing

........

The premium Read AI charges over Fathom and Fireflies is a direct trade for Search Copilot's cross-communication reach and bundled premium-model access included from the Pro tier rather than an Enterprise add-on — a real justification for organizations that specifically want Slack- and email-integrated search, but not one that pays off for a team whose need stops at meeting transcription and summaries alone.

··········

THE DECISION RULE FOR EVALUATING READ AI.

Read AI earns its price specifically for organizations that want a unified search layer across meetings, email, and chat rather than isolated meeting summaries — Search Copilot's cross-communication reach, combined with included GPT-4.1 and Claude access, is the feature that justifies the premium over Fathom's free-tier generosity. A team whose need is purely meeting transcription and summaries with no interest in engagement scoring or cross-app search gets more value per dollar from Fathom's unlimited free tier. Any European organization considering Read AI specifically for its Engagement Score and sentiment-analysis capability should verify current EU AI Act compliance directly with the vendor before deployment, given at least one detailed review's finding that the feature doesn't currently meet that standard, and any team relying on the Charisma coaching feature specifically should test it directly rather than assume it delivers actionable feedback based on the product's broader feature list.

·····

FOLLOW US FOR MORE.

·····

DATA STUDIOS

·····

datastudios.org