Last week, I witnessed the future of job interviews. After clicking through to a web browser, a disembodied voice welcomed me with a confident, conversational tone indistinguishable from a human’s. “Hi, Leo,” she said. “I’m AIR, and I’ll be conducting your interview for the designer role at Reddit.”
For the next 10 minutes, AIR probed me on my (fictitious) experience, pivoting based on my responses and incorporating my answers into her follow-ups. AIR, of course, was an AI-powered voice bot, but seemed no different from a hiring manager conducting a Zoom interview with her video off, except for a few small hiccups. At the end, I received a score of 4.6 out of five and feedback that I would be “a valuable addition to any design team.” (Unfortunately, not true, but that’s not AIR’s fault.)
The experience was not lifelike enough to evoke the so-called uncanny valley, but it still offered spooky confirmation that AI is ready to disrupt the hiring process. Or at least that’s what recruiting platform Braintrust hopes to prove with the launch of AIR on Monday, which will go live for a number of high-profile clients, including Nestlé and Guardian Life, enabling hiring managers to drastically cut time and money spent on interview pipelines through AI-led, autonomous video screenings.
“This is doing to job recruiters what Kayak did to travel agents,” Braintrust cofounder Adam Jackson said in an interview with Fortune, describing the new product as a “classic 2024 story of AI totally replacing a major role in the economy.”
Crypto meets AI
Founded in 2018, Braintrust has long been on a mission to use emerging tech to rethink the job search. For the past six years, blockchain has been at the core of Braintrust’s product, with the company launching its own Ethereum token (also called Braintrust) as an incentive mechanism and payment system to connect recruiters, candidates, and hiring managers, promising a lower take rate than competitors like Fiverr or even LinkedIn. In late 2021, Braintrust announced that the investment giants Tiger Global and Coatue had purchased $100 million worth of its token.
Braintrust’s crypto journey hasn’t always been a smooth one. As tech critics like Charlie Warzel have pointed out, Braintrust’s hiring marketplace did not necessarily need a token at its core, although it did create a new store of value and rewards tool that the platform did not need to pay for itself. (A Braintrust token is currently priced at about 80 cents, with a market cap of over $175 million.) The crypto aspect has also created complications, including Tiger selling large tranches of its token holdings and contributing to the price of Braintrust slumping over 20%.
The organization argues that the token creates a more “decentralized” system that allows network participants, from job candidates to recruiters, to participate in governance, as well as reducing fees associated with middlemen like staffing agencies. Still, Braintrust operates largely like other job platforms, with companies from Deloitte to Nike posting open positions, and recruiters recommending candidates, which are both rewarded for successful matches. The platform’s jobs are concentrated in engineering, design, and marketing, and range from freelance to full-time. According to Jackson, Braintrust has a pool of approximately 750,000 vetted candidates serving 2,500 companies.
Braintrust has also been exploring AI for the past year and a half, including building its own large language model, or LLM, atop Meta’s Llama 2. The company’s first experiment involved automating the process of matching clients’ roles with applicants, a feature that went live about six months ago. Through the tool, a hiring manager can enter keywords for the type of role they’re hiring for, such as a mobile developer, and Braintrust will populate a tailored job description and post the listing on its platform, as well as other job boards, including LinkedIn, and filter through applicants.
The new feature—called AIR, or AI Recruiter—will go a step further. As I experienced, AIR will schedule and conduct interviews with the top candidates, asking questions based on clients’ specifications through a live, conversational process. At the end, AIR creates a scorecard and recommends the top applicants, with hiring managers able to go back and watch the interview.
According to Jackson, AIR crunches a screening process that normally takes hundreds of hours and tens of thousands of dollars into “zero dollars in 24 hours.”
Though the feature has about 20 clients signed on for the launch next week, AIR is still untested outside Braintrust, with my mock interview representing the first one conducted by a nonemployee.
Still, as crypto entrepreneurs search for ways to integrate the buzzy field of AI into their products, Jackson hopes that AIR will represent one of the first tangible use cases. “We give you good reasons to earn the token,” he said. In one scenario, job hunters could earn Braintrust tokens by referring other people to use the network, and then use their rewards to conduct mock interviews with AIR, which is a feature Jackson aims to roll out soon.
“When you’re earning Braintrust tokens, you now own a piece of the network,” he told Fortune. “The more the network grows, and hopefully the value of the token grows, you’re an owner now, whereas with cash you’re not an owner.”
