Meetings take time, and there's no way around that. According to his 2022 poll from Deputy.com, the majority of U.S. employees spend up to about eight hours in meetings each week, depending on industry and region.
Reduced productivity explains the growing popularity of AI-powered summarization tools. In a recent survey of marketers conducted by the Conference Board, a nonprofit think tank, nearly half of respondents said they use AI to summarize the content of emails, conference calls, etc. I answered.
Many video conferencing suites now have summarization built-in, but David Shim believes there is room for third-party solutions. He is the co-founder of his Read AI, which brings together video calls across platforms such as Zoom, Microsoft Teams, and Google Meet.
Shim, previously CEO of Foursquare, co-founded Read AI in 2021 with Rob Williams and Elliott Waldron. Prior to Read AI, the trio worked together on his Foursquare, Snapchat, and Shim's previous startup, Placed (which Foursquare acquired his in 2019).
“Read AI's direct competition is traditional project management, where notes are written by hand,” Shim told TechCrunch. “By learning what's important to you cross-platform, Read is not your co-pilot, but rather your autopilot, delivering content that makes your work more effective and efficient.”
Initially, Read focused solely on video conferencing solutions, offering dashboards to measure meeting progress (at least as determined by certain metrics) and two-minute summaries of hour-long meetings. But coinciding with his recently completed $21 million funding round led by Goodwater Capital and in collaboration with Madrona Venture Group, the company is also expanding into message and email summaries. Masu.
Read's new features are available in a “soft launch” and allow you to connect to Gmail, Outlook, and Slack, as well as video conferencing platforms, to learn about topics that may be relevant to you. Within 24 hours of connecting to your messaging or video conferencing service, Read delivers daily summaries, AI-generated “takeaways,” summaries of key content, and updates on conversation topics in chronological order. start. Read charges between $15 and $30 per month for its services.
“What’s unique about Read is that its AI agent works silently in the background and allows meetings, emails, and messages to interact with each other,” Shim said, adding that an average summary from Read AI A summary of 50 emails from 10 people condensed into one. “This connected intelligence unifies communications and enables us to deliver actionable briefings that are personalized to your needs and priorities.”
Now, call me a skeptic, but I'm not sure I can trust an AI-powered tool to summarize content consistently and accurately.
Read's platform leverages generative AI to summarize meetings, messages, and emails.Image credit: Read
Models like ChatGPT and Microsoft's Copilot are prone to illusions and make mistakes when summarizing, such as in meeting summaries. In a recent article, the Wall Street Journal described an early adopter using Copilot for meetings with an example where Copilot fabricated attendees and implied that the calls were about topics that were never actually discussed. I am listing it.
Are Read AI's tools any different? Shim claims it's more robust than many solutions out there, including rivals like Supernormal and Otter.
“Read implements a unique methodology that adjusts the raw content and language model output, so deviations are automatically detected and appropriately controlled,” he said. “Additionally, meeting content can be used to better contextualize email and messaging content, further reducing uncertainty and improving outcomes.”
Please take this statement with a grain of salt. Sim did not share benchmark results to support these claims.
Shim emphasized that instead of benchmarking, summarization tools like Read can (in theory) deliver productivity gains.
“Instead of you being late or double-booking and rescheduling meetings, Mr. Reed will attend on your behalf and provide you with an overview and action items that even the best executive assistants can't match. He said, and Reed also emphasized. He doesn't use customer data to train his AI models, and users have “full control” over the content that passes through the platform. “AI is refocusing knowledge workers.” [by] It saves you hours a day. ”
Read AI is full of controversy, so it's a little hard to take Sim at his word. The platform's sentiment analysis tools, which interpret meeting attendees' voices and facial cues to inform organizers about sentiment, have been criticized by privacy groups as being overly invasive, prone to bias, and that the data It has been criticized for being a very high security risk.
Gender and racial bias are well-documented phenomena in sentiment analysis algorithms.
Sentiment analysis models tend to assign more negative emotions to Black faces than white faces, and perceive language used by some Black people as offensive or harmful. AI video recruiting platforms have been shown to respond differently to the same job candidate wearing different clothing, such as glasses or a scarf. Also, his 2020 research at MIT showed that algorithms can be biased towards certain facial expressions, such as smiling, reducing accuracy.
Image credit: Read
Presumably, while Sim continues to view Read's sentiment analysis technology as a competitive advantage rather than a risk, customers can disable this feature and analytical data is periodically deleted from Read's servers. It is pointed out that it will be done.
“Using a multimodal model allows Reed to incorporate nonverbal responses into the meeting outline,” he said. “As an example, a startup company might talk about the benefits of its product during a pitch meeting, but participants visibly shake their heads or frown during the pitch…Read is Create custom baselines of engagement and sentiment for your users, apply a one-size-fits-all model, and ensure that each person is treated as a unique human being.”
Accurately or not, some are confident that Read, with a $32 million war chest and a customer base that has grown by half a million users in the past quarter, can clearly deliver on that promise. There is.
Reed, based in Seattle, Washington, plans to use the new capital injection to double its workforce to more than 40 by the end of the year, Sim said.
“In the face of a widespread slowdown over the past few years, Mr. Reed has continually seen the growth curve steepen across users, meetings and revenue,” he added. “This accelerated growth can be directly attributed to the quantitative benefits users receive in terms of time savings when using Read AI in meetings.”