While working at Uber and Palantir, respectively, Siadhal Magos and Shahriar Tajbakhsh noticed that the hiring process, especially the interview process, had become unwieldy for many companies' human resources departments.
“While the interview was clearly the most important part of the hiring process, it was also the most opaque and unreliable part,” Magos told TechCrunch. “On top of this, taking notes and writing feedback is a lot of effort, which many interviewers and hiring managers try to avoid at all costs.”
Magos and Tajbakhsh believed the hiring process was ripe for disruption, but they wanted to avoid over-abstracting the human element. So they launched Metaview, an AI-powered note-taking app for recruiters and hiring managers that records, analyzes and summarizes job interviews.
“Metaview is an AI note-taker built specifically for the hiring process,” Magos says. “This allows recruiters and hiring managers to focus on getting to know the candidate rather than extracting data from conversations. It will save you a lot of time writing and you will be able to focus more during the interview because you won’t have to multitask.”
Metaview integrates with apps, phone systems, video conferencing platforms, and tools like Calendly and GoodTime to automatically capture interview content. Magos said the platform will “account for the nuances of hiring conversations” and will be “enhanced with data from other sources” such as applicant tracking systems to highlight the most relevant moments. ing.
“Zoom, Microsoft Teams, and Google Meet all have transcription built-in and are potential alternatives to Metaview,” Magos said. “However, the information that Metaview’s AI pulls from interviews is far more relevant to the hiring use case than the generic alternatives, and we also use tools to improve users’ hiring workflows within and around these conversations. We will help you take your next step.”
Image credit: Metaview
It's true that traditional job interviews have a lot going wrong, and note-taking and conversation analysis apps like Metaview could help, at least in theory. According to an article in Psychology Today, the human brain is full of biases that interfere with judgment and decision-making, such as over-reliance on the initial information presented or in a way that confirms pre-existing beliefs. This includes the tendency to interpret information. .
The question is, does Metaview work, and more importantly, does it work equally well for all users?
Even the best AI-powered voice dictation systems suffer from their own biases. A Stanford University study shows that black speakers' error rates in speech-to-text services from Amazon, Apple, Google, IBM, and Microsoft are nearly twice as high as their white counterparts. Another more recent study published in the journal Computer Speech and Language found statistically significant differences in how the two leading speech recognition models handle speakers of different genders, ages, and accents. I did.
Hallucinations also need to be considered. AI makes mistakes when summarizing, such as meeting summaries. In a recent article, the Wall Street Journal reported on an early adopter using Microsoft's AI Copilot tool to summarize meetings, where Copilot fabricated attendees and implied calls about topics that were never discussed. I'll give you an example of what happened.
When asked what steps Metaview has taken to mitigate bias and other algorithmic issues, Magos said that Metaview's training data “exceeds human performance” in recruiting workflows. They argued that there was enough diversity to produce a model that performed well on common benchmarks for bias.
I'm skeptical and a little wary of Metaview's approach to how it handles audio data. According to Magos, Metaview stores conversation data for two years by default unless the user requests that the data be deleted. That seems like a very long time, and probably the candidates too.
However, none of this appears to have affected Metaview's ability to raise capital or acquire customers.
Metaview raised $7 million this month from investors including Plural, Coelius Capital and Vertex Ventures, bringing the London-based startup's total funding to $14 million. Magos said Metaview's customer base has grown 2,000% year over year to his 500 companies, including Brex, Quora, Pleo and Improbable.
“This funding will be used primarily to grow our product and engineering teams, and will further fuel our sales and marketing efforts,” Magos said. “We've tripled our product and engineering teams, further fine-tuned our conversation synthesis engine to ensure our AI automatically extracts exactly the right information customers need, and we're also working to eliminate inconsistencies in the interview process and candidates who seem Develop a system that proactively detects problems such as problems that cause you to lose interest.”