Imagine you are looking for a new job and find a role you like listed on an employment website.
As per usual, you send in your CV, emailing it to the contact provided.
Typically, the next step would either be a polite "no thank you" email or a message to contact a recruitment professional or HR department. In some cases, you might never hear back (in which case you understand you were nowhere near the mark).
Companies and recruitment agencies using the solution provided by Sydney startup Curious Thing, in contrast, have a very different experience.
Everyone who sends in his or her CV receives a quick email asking them to call a specific phone number. Please call in three days, at a time convenient to you, if you are still interested in the position.
When you call, you are interviewed. But not by a human being. Your conversation and the entire interview is conducted by artificial intelligence.
“We believe that conversational AI and natural language processing will essentially help us to collect more information in a nontraditional way,” said Sam Zheng, one of three co-founders at Curious Thing, which was founded in July 2018.
“With more data points and information from the interview, it can be particularly helpful for improving the recruitment process.”
Zheng makes the point that around 90% of all CVs submitted for jobs are barely even skimmed over. But people are rejected or accepted mainly based on their CVs, which are not benchmarkable.
They go through a very cursory process of selection to create a short list. The process is fundamentally flawed not only by the time pressures involved but also by inherent biases.
HR recruiters can only make a certain number of calls in a day, and it is almost a physical impossibility to interview everyone who applies.
In terms of bias, graduates from certain universities or schools can be subtly discriminated against because these institutions don’t have the status of others. In these cases, discrimination is against the institution and has very little to do with the person.
With Curious Thing, every person who applies for a position is interviewed by the AI. The information is then presented in a benchmarked and structured way.
"Most companies spend around 70% of their time in the pre-interview phase plus the CV screening. The most critical decisions are made upon shortlisting,” said Zheng.
“We are providing more data points around experience, qualifications and motivation. We are fundamentally trying to help people decide when they go down from 200 candidates to 10. We make sure they are talking to the right ten."
Zheng said Curious Thing is not recruitment by AI, with the AI making the decisions. Instead, it is humans working with the AI to save time, be more thorough, and eliminate unconscious bias.
“By taking humans out of the equation we are trying to help humans regain focus,” he said.
"AI is not trying to find which person is better. The AI is trying to ask the questions based on our clients required attributes. After each conversation with the AI, we assess each particular person against these specific requirements.”
While the specific requirements are built into the AI’s model, what is rejected are profiles of previous employees hired for similar roles by the client, as these selections might have also involved some kind of bias.
Self-serve on the Horizon
It is early days for Curious Thing. But the startup has already caught the attention of venture capital and secured AUD 1.5 million in seed funding in February 2019 from Singapore-based Qualgro, and Australian VC Reinventure, which is part of Westpac Bank.
Zhang, who was originally an actuary by profession and training, is himself something of a serial startup entrepreneur. He founded Australian business analytics startup Hyper Anna, which is also part of the Westpac Reinventure portfolio.
In its brief life so far Curious Thing has been adopted by corporates, though it has yet to penetrate into the specialized HR recruitment market.
On the horizon, said Zheng, is a “self serve tool” which clients will be able to configure and deploy themselves and access through a software-as-a-service model.
It will make AI recruiting affordable and accessible to smaller and mid-market companies, who might now be missing out on significant talent simply because they don’t have the time or resources to screen all the CVs they get when they post jobs.