“When I moved into HR, I was immediately surprised by the complete lack of data on people’s decisions,” she says.
It made no sense to her, especially given how BCG prioritizes data and evidence in its client work.
“When you’re advising the board on growth strategies, cost reductions, or any kind of change, you have to bring data into the conversation, otherwise it’s really unreliable… Aside from books, what I can argue is that when people look at data, they make decisions entirely on intuition.”
The head of the REA Group’s human resources and cultural departments has caused the absurdity of the house for Hyman. Relying on subjectivity in hiring decisions meant that the company was only hiring people with the same engineering or start-up backgrounds as its existing staff, which quickly became an asset to the company’s development. became a problem.
“By reinforcing biases when hiring, we miss out on a lot of talent, which is costly to earn, retain and retain. [and] Innovation is usually invisible, but it can be significant,” says Hyman.
Extensive academic research has shown that no amount of human effort can overcome unconscious biases. Many technical tools can do neither, as programmers have built their own biases into their functionality.
But AI has AI potential, and in this regard Hyman finds it a “realization” of her vision of improving how we prioritize both data and diversity in hiring decisions. Did.
Like an interview, but different
Sapia.ai sorts applications and selects shortlists for interviews. The interview will be conducted via the “smart chat” option. This is essentially a chatbot. It takes 20-25 minutes to complete, but applicants can choose when and where to do it.
Sapia.ai selects the best applicants (or shortlists) from these interviews and hands them over to the client. Hyman says this final step ensures that “humans are still involved.”
Not only is it programmed with traditional bias pitfalls (gender, race, class, education, work history), but AI can also help determine whether applicants are best suited for the role, rather than, say, most confident or likeable in an interview. It also tests whether it shows . .
“It’s designed to simulate an interview, but without the tension of an interview,” she says. “You don’t have to show your face. We’re not pressed for time.”
Sapia.ai’s strength and global competitive advantage lies in its breadth of data. The company’s proprietary dataset contains his 1 billion words extracted from his 12 million interview questions from 2.5 million users. This is also first-party data, unlike the third-party data collected from his resume and his LinkedIn by rivals.
This data powers Sapia.ai’s “smart chat” functionality. The feature is populated only with language from interviews, and is therefore “almost blind to the prejudices” that would have been ingrained in the program or review by humans.
Recruiters are also trained to recognize the soft and hard skills required for a role, and to detect plagiarism, anomalous answers, and, ironically, AI-generated interview answers.
“I don’t know of any other company that allows you to discover a person through a short conversation, which is very interesting considering we are a small Australian company,” says Hyman.
One of Hyman’s main reasons for entering the US market is that he wants more data.
“It’s always been part of our business strategy to focus on opportunities that give us a lot of data, because that’s our power,” she says.
“But there aren’t enough companies like Woolworths in Australia operating at scale to really help with our datasets, so going to the US was a natural step.”
Diversity Tailwind
Another reason why this ‘small Australian company’ is stepping onto the world stage comes down to its second objective: diversity.
Hyman’s picky approach to building Sapia.ai’s customer list seems at odds with how capital- and customer-hungry start-ups usually operate.
“This might sound really crazy for a startup, but we’ve been pretty selective about the companies we target,” she says.
She only wants to partner with “early adopters or fast followers.” Taking Woolworths as an example, Hyman believes the company has “the most impressive HR team in the country.”
“We want customers who also have a deep passion for diversity and the candidate experience… you will find someone who understands that and that really increases your credibility. Our brands are actually built from our customers’ brands.”
In other words, you have to walk the talk, and Hyman thinks you’re more likely to find those kindred spirits in the United States.
“Diversity is important here, but it doesn’t get us a deal. If you can, the world will be on your side.”
The launch in the U.S. is still in its early stages – Hyman flew there from Melbourne just after our interview to kick the business up – and the CEO and founder says she believes she He admits that he is “never happy and always anxious.”
The overseas market is usually “more merciless” than the Australian market, she worries, and launching there “can quickly run out of money”.
But she still hopes that Sapia.ai’s passion for data and diversity will help her compete in an environment littered with tech peers and AI-obsessed people.
“The United States is a country built on innovation. Australia tends to be more conservative, but if we can [clients’] If you solve the problem, it actually works, and they trust you the way you do it, then you have a chance. So now I just have to find that opportunity.”
