Meetings are time-consuming and there isn’t a way around it. According to a 2022 survey by, there are a lot of U.S. employees spend as much as eight hours in meetings each weekdepending on the industry and site.

The decline in productivity explains the growing popularity of AI-powered summarization tools. In a recent survey of marketers by The Conference Board, a nonprofit think tank, Almost half of those surveyed said they used AI for summarization the content of emails, conference calls and more.

While various video conferencing suites now offer built-in summarizing capabilities, David Shim believes there’s room for third-party solutions. And he would: He is a co-founder of Read AIwhich aggregates video calls on platforms similar to Zoom, Microsoft Teams and Google Meet.

Shim, previously CEO of Foursquare, co-founded Read AI in 2021 together with Rob Williams and Elliott Waldron. Before Read AI, the trio worked together at Foursquare, Snapchat and Shim’s previous startup Placed (which acquired Foursquare in 2019).

“Read AI’s direct competition is traditional project management, where notes are written manually,” Shim told TechCrunch. “By learning what matters to you across platforms, Read just isn’t a co-pilot but an autopilot, delivering content that makes your work more practical and efficient.”

Read initially focused solely on video conferencing solutions, offering dashboards for measuring how a gathering was progressing (at the very least based on certain metrics) in addition to two-minute summaries of hour-long meetings. But coinciding with a recently accomplished $21 million funding round led by Goodwater Capital and Madrona Venture Group, the corporate is expanding messaging and email aggregation.

Available in soft launch, Read’s latest feature connects to Gmail, Outlook and Slack, in addition to video conferencing platforms, to learn topics that could be relevant to you. Within 24 hours of connecting to the messaging and video conferencing services you employ, Read will begin delivering each day updates with summaries, AI-generated takeaways, an outline of key content, and updates on talking points in chronological order. Read charges a monthly fee of $15 to $30 for its service.

“What makes Read unique is that its AI agents work quietly within the background, allowing your meetings, emails and messages to interact with one another,” Shim said, adding that Read AI’s average summary is 50 emails. Emails from 10 recipients summarized in a single summary. “This connected intelligence unifies your communications and empowers you and your team with personalized, actionable briefs tailored to your needs and priorities.”

Now let me be skeptical, but I’m undecided I trust an AI-driven tool to consistently summarize content accurately.

Read’s platform uses generative AI to aggregate meetings, messages and emails. Photo credit: To read

Models like ChatGPT and Microsoft’s Copilot Make mistakes when summarizing attributable to their tendency to hallucinate, also in Session summaries. In a recent article within the Wall Street Journal quoted A case where Copilot, for an early adopter using Copilot for meetings, invented participants and implied that the calls were about topics that were never actually discussed.

Is Read AI’s tool different? Shim claims it’s more robust than lots of the solutions available, including competitors like Supernormal and Otter.

“Read runs a proprietary methodology to coordinate raw content with language model output in order that deviations are robotically detected and controlled accordingly,” he said. “In addition, we are able to leverage content from meetings to higher contextualize email and messaging content, further reducing uncertainty and improving outcomes.”

Take this statement with a grain of salt. Shim has not published any benchmark results to support these claims.

Instead of benchmarks, Shim emphasized productivity-enhancing summarization tools like Read (in theory).

“Instead of rescheduling a gathering since you’re late or double-booked, Read can attend in your home and give you a summary and motion items that even the most effective executive assistant couldn’t handle,” he said, also emphasizing that Read uses doesn’t use customer data to coach its AI models and ensures that users have “full control” over the content transmitted through the platform. “AI brings knowledge employees back into focus by saving them hours within the day.”

Read AI is not any stranger to controversy, so it’s hard to take Shim at his word. The platform’s sentiment evaluation tool, which interprets the vocal and facial signals of meeting participants to tell hosts about their mood, was developed called challenged by data protection advocates overly invasive, susceptible to bias and potentially a risk to data security.

Gender and race Prejudices Area Sodocumented phenomenon In Feeling evaluation Algorithms.

Emotional evaluation models are likely to assign more negative emotions to the faces of black people than to those of white people, and perceive The language some black people use is taken into account aggressive or toxic. AI video hiring platforms were found reacting otherwise to the identical applicant wearing different outfits similar to glasses and a headband. And in 2020 study MIT researchers showed that algorithms could possibly be biased towards specific facial expressions, similar to smiling, which could affect their accuracy.

Read AI

Photo credit: To read

Significantly, Shim continues to view Read’s sentiment evaluation technology as a risk slightly than a risk, but notes that customers can opt out of the feature and that analytics data is periodically deleted from Read’s servers. “By using a multimodal model, Read can incorporate nonverbal responses into meeting summaries,” he said. “For example, during a pitch meeting, a startup is likely to be talking concerning the advantages of the product, but participants are visually shaking their heads and frowning throughout the pitch… Read creates a custom engagement and mood baseline for every meeting participant as a substitute of applying a one-size-fits-all model that ensures everybody is treated as a singular individual.”

Whether true or not, with a war chest of $32 million and a customer base that grew by half one million users within the last quarter, Read has clearly convinced some people who the corporate can deliver on its guarantees.

Read, based in Seattle, Washington, plans to double its workforce to greater than 40 employees by the top of the yr to reap the benefits of the brand new capital infusion, Shim said.

“Amid an overall slowdown in recent times, Read has continued to see the expansion curve in users, meetings and revenue steepen,” he added. “This acceleration in growth might be directly attributed to the quantifiable return on investment that users realize in the shape of time savings from using Read AI of their meetings.”

This article was originally published at