What’s next for AI-enabled recruiting and talent acquisition?

Long-retired Clippy sought to help users navigate Microsoft Word—popping up uninvited, blinking aggressively, and occasionally even tapping on the monitor glass.

A future might not be far off wherein generative AI assistants help HR professionals navigate complicated application tracking systems (ATS), and—provided Clippy’s AI-powered descendents are more useful, and respectful of personal space—this could prove to be a boon for talent professionals looking to streamline their craft.

“[Generative AI is] evolving so fast, and I’m very encouraged by some of the companies in the market… thinking about, how can this technology become responsible AI by being more explainable? It’s not there yet,” iCIMS CTO Al Smith said.

Because ATS tools are a “different class of software” than most others used in the industry, Smith said that developers are designing and testing solutions to ensure recommendations are accurate and add value to the TA professional.

“If you really distill it down, [ATS software is] a 30-year industry of taking what had been a paper process, 30 years ago, and digitizing it, and so there’s still an immense amount of content creation, whether that’s job description, marketing campaigns, offer letters,” Smith said. “We see huge value where this technology helps generate a higher level and higher quality starting point for you.”

The company released in March two AI features that are “relatively simple,” said Smith, with its Copilot tool helping HR clients address some of the myriad content creation associated with recruiting and hiring.

The AI, which is informed by iCIMS’s database of job descriptions, can be used to help recruiters or hiring managers draft job descriptions that call for the right skills and experiences. It’s also “particularly helpful” if a company needs to reimagine a job description as more senior or junior, he added.

iCIMS also designed a tool that relies on generative AI to help create custom interview questions based on a job description, as well as relevant information related to a candidate. Smith said the team at iCIMS asked the models to provide users with expected answers and explain why it suggested the questions.

Smith identified eight areas where the technology could be useful to TA pros, and “there is an underlying theme that it deals with content of some form.”

“One thing AI is always good at is solving problems beyond human scale, when you have too much of something,” he said. “So those are really high priorities in our use cases…What we’re not doing at this point is using generative AI to do job match, people match selection.”

Beyond the content

One intuitive place where AI might come in handy is linking candidates to jobs, but for iCIMS (and other ATS software companies), this is a thorny space legally and ethically.

“We started our program with a code of conduct and ethics for AI,” Smith said. “It’s human-led, it’s explainable, and it’s transparent and reversible. [In its] current state, generative AI is none of those things.”

When it comes to assisting HR teams in hiring decisions, the technology doesn’t meet the company’s standards, and making the right and ethical hiring decision is too critical a component of the business—especially in today’s labor market—to get it wrong.

“How important is [it] that [AI-generated] artwork is correct? It’s not; it’s a subjective decision. AI is fine for things like that,” he said. “But when precision and regulatory rules are required, generative AI is not there yet.”

It’s also on iCIMS to make clients sure using its software won’t run afoul of new compliance measures aimed at protecting against bias in the hiring process. Smith sees AI as a “complement [to] that selection process” and provides TA pros with materials related to candidates that might be useful when considering a hiring decision.

“The program that we have in place, we work so hard at…curating data to make sure that we’re not using data that we know creates bias,” he said.

Consulting an AI-xpert

Clippy was excellent at knowing how to use Microsoft Word…he just failed to realize that many of its users knew how to use it, too.

But Smith sees an opening for an AI assistant to help navigate the complicated options and processes when interfacing with the iCIMS platform. An AI assistant could serve TA professionals with institutional knowledge about how a company has done one thing or another in the past and nudge users to remember a step they might have skipped. This, Smith said, is even more important for hiring managers or HR professionals who may have taken over TA roles following widespread recruiter layoffs in recent years.

“Software is hard to use, especially business software,” he said. “I find that it’s hard for people to be an expert on using our software, and I do see the opportunity…that the Copilot experience can actually be the expert user.”