Employers deploy AI-powered chatbots to screen job applicants

Job applicants who need to get passed that first, all-important screening interview may soon find themselves face-to-face with a robot.

Chatbots and intelligent software to analyze video interview answers are becoming practical methods of screening applicants at scale, particularly for those industries with high volume staffing needs such as retail, hospitality, and call centers.

Three cases in point: Marriott introduced a Marriott Careers Facebook Messenger chatbot in November. Unilever is making use of artificial intelligence to analyze video interviews. HotTopic relies on AI to screen applications and plans to begin experimenting with AI-powered chatbots next year.

Marriott Careers takes advantage of Google cloud services for natural language processing and sentiment analysis, says Svjetlana Vukic, head of client service at Five, the digital agency that created the chatbot for the hotel chain. It also builds on Dexter, a software platform that acts like a content management system for a library of questions and answers.

marriott
Palm trees stand in front of the Marina Del Rey Marriott hotel in Marina Del Rey, California, U.S., on Monday, March 21, 2016. Starwood Hotels & Resorts Worldwide Inc. accepted an improved takeover offer from Marriott International Inc. valued at $13.6 billion, proceeding with plans to form the world's largest hotel operator after investors led by China's Anbang Insurance Group Co. sought to thwart the deal. Photographer: Patrick T. Fallon/Bloomberg
Patrick T. Fallon/Bloomberg

While the chatbot may not understand everything a job candidate says, it is set up to answer common questions like, “Where are the best bartending jobs in Chicago?” and “How do I get in contact with a recruiter?” On the back end, the software taps into Oracle’s Taleo talent management system to look up openings that match those job categories and locales.

The Marriott chatbot is not really AI—or at least not as advanced an application of AI as the customer-facing chatbots that Marriott has introduced—and it doesn’t really learn as it goes. The software is powered, however, by data-driven logic and specialized algorithms and, according to a company representative, may be packed with more AI smarts in the future.

Overcoming biases
In the case of Hot Topic, “At the end of the day, we want to hire the most qualified person for the role,” says Johnny Sanchez, head of recruiting at the retailer. Given the volume of hires his stores must make and the number of applicants they must sort through, the “conscious or unconscious biases” of hiring managers can get in the way. Store managers, he says, may choose a candidate because they went to the same school or base their decision on purely subjective preferences, like whether they like the person’s name.

To correct for this, the retailer makes use of artificial intelligence from Ideal, a software developer that markets a virtual assistant for job recruiters. The AI-driven program helps the retailer review more applications, more thoroughly and objectively, Sanchez says.

The Ideal platform screens every resume using the same, consistent criteria, he explains. It also incorporates feedback from Hot Topic’s applicant tracking system to continually refine its best hire model.

When the retailer first deployed the AI application, connecting it to a chatbot struck Sanchez as “too bleeding edge.” But this coming year he plans to pilot an intelligent chatbot with Ideal.

Going beyond textual analysis, Unilever has put AI to work analyzing the video interview responses job candidates give to a series of scripted questions. First, candidates complete a set of job aptitude assessment games from Pymetrics, a developer of AI-driven recruitment software, which steer them toward the most suitable jobs. Then, using technology from HireVue, Unilever records video interview answers to a set of scripted questions.

The HireVue software uses AI to tease meaning from the candidates’ answers, including meanings conveyed by tone of voice and facial expression. If an applicant says how much she likes her current boss, but her tone of voice doesn’t match what she’s saying, the software will note the discrepancy.

HireVue’s biggest advantage, according to the company’s chief technology officer Loren Larsen, is that while a human recruiter might be limited to conducting 10 video interviews a day, the software can review thousands—and then forward the most promising ones to a recruiter for further review. One way that Unilever has benefitted from this, Larsen says, is that previously its college recruiting program could only cover a limited number of schools.

Per a case study published by Bersin by Deloitte, Unilever’s AI initiatives have saved the company more than 50,000 work hours and erased months off the time it took to hire 800 candidates for the company’s Future Leaders Programme.

For reprint and licensing requests for this article, click here.
Recruiting tools Recruiting HR Technology Employee engagement Benefit communication
MORE FROM EMPLOYEE BENEFIT NEWS