How micro1’s AI interviewer could make tech hiring more efficient and fair

Don’t miss OpenAI, Chevron, Nvidia, Kaiser Permanente, and Capital One leaders only at VentureBeat Transform 2024. Gain essential insights about GenAI and expand your network at this exclusive three day event. Learn More

Micro1, a startup applying artificial intelligence to recruiting that we covered last year, announced the launch of an AI-powered technical interviewer aimed at helping companies efficiently screen software engineering candidates at scale. The tool generates tailored questions based on candidates’ self-reported skills, conducts voice-based technical interviews and coding assessments, and produces detailed evaluation reports.

In an exclusive interview with VentureBeat, micro1 founder and CEO Ali Ansari explained that the product targets a key pain point in tech hiring. “Usually when you do a job posting, especially with global [reach], you get many, many applications,” Ansari said. “Employers usually just pick a random sample size from those applications and interview that. The decision on picking which ones to interview is very arbitrary.”

Customized questions and real-time assessment reports

Micro1’s AI interviewer aims to solve this by enabling companies to consistently evaluate a much larger portion of applicants. Candidates input their top skills and seniority level for each, such as “senior” for React. The system then dynamically generates questions testing relevant theoretical and practical knowledge. 

Ansari emphasized the focus on customization and real-time content generation. “The questions that we ask are based off [the skills input], so if someone puts in React and Node.js as their top skills, we’re gonna ask them theoretical programming questions on React and Node.js,” he explained. “There is no static database of questions. We generate questions in real-time using a language model, so there’s new questions that are randomized each time.”

Countdown to VB Transform 2024

Join enterprise leaders in San Francisco from July 9 to 11 for our flagship AI event. Connect with peers, explore the opportunities and challenges of Generative AI, and learn how to integrate AI applications into your industry. Register Now

After completing a voice-based Q&A and live coding, candidates receive an automatically generated assessment report that rates performance in each skill area. This report helps recruiters prioritize top performers to advance to human interviews. Micro1 reports that this screening has increased the pass rate for human interviews to around 50%, compared to a typical 10-15%.

Addressing AI bias and candidate experience

The young company is keenly aware of industry concerns around AI bias. While acknowledging that some bias is inevitable, Ansari highlighted multiple steps taken to mitigate unfairness. Questions are refined based on human interviews, and “the AI interviewer cannot say this candidate passed or failed, that doesn’t exist.” Recruiters must holistically consider the AI assessment alongside other data points like resumes.

Reactions from early candidates have been largely positive, with Ansari estimating 80-90% very favorable feedback. “A lot of [responses are] like ‘wow, this was actually really enjoyable, I was less nervous, I was able to articulate my thoughts better,’” he said. Over time, the company aims to make the AI experience “on par” with or “even better” than human interviews.

Differentiating in a growing AI recruiting market

Micro1 is part of a growing wave of startups building AI tools for recruiting. Competitors like Filtered and Karat also offer automated technical assessments. But micro1 differentiates through its dynamic interviews and aim of “giv[ing] qualified candidates a better shot by having them actually convey their skills,” in Ansari’s words.

As companies struggle to fill nearly 400,000 open computing jobs in the US alone, tools to optimize hiring are in high demand. micro1 has already signed a number of early clients, primarily midsize tech companies feeling the talent crunch acutely. But Ansari is bullish about expanding to additional roles and company sizes long-term.

The product’s launch comes amid a fierce debate around the appropriate use of AI in high-stakes domains like hiring. Opponents argue that today’s AI systems are too unreliable and biased to make fair, consequential assessments about humans. Proponents counter that when used carefully, AI can help reduce human bias and missed opportunities at scale.

Ansari and the micro1 team firmly believe that, with responsible implementation focused on augmenting rather than replacing human decision-making, AI interviewers can make hiring both more efficient and equitable. “The goal is to get the bias in these AI systems to be less than humans in this use case,” Ansari said. Time will tell if micro1 can deliver on that ambitious vision.