formulating the job description

Once again, the basic elements of the job description need to come from internal experts at your company. Only you and your colleagues know details like the job’s salary, key responsibilities and required qualifications — although using generative AI tools as a source of interactive feedback could be useful.

Generative AI really comes in handy when it’s time to package these basic details into an engaging, attractive job description that attracts top talent. Hand the details over to a generative AI chat tool and give it some guidance on tone of voice and other requirements like inclusive language, and it’ll instantly create an AI-generated job description. The output will require analysis and tweaking, but using AI in recruitment in this way can improve the speed of creation and the quality of your job descriptions — especially if the recruitment team is short on writing skills.

Generative AI can also be used to perform job analysis and help identify skills, education, and experience that would be useful in the role, expanding your talent pool. It can also be useful to leverage generative AI tools to help create outcome-based job descriptions, which can drive up the quality and quantity of candidates. Both of these use cases can be performed by recruiters, the HR team, and hiring managers.

searching and matching talent

This is the stage where recruiters distribute the job on relevant platforms, reach out to their own candidate networks, and start receiving applications. General-purpose, free-to-use generative AI tools might be tempting to use in this stage, but consent and privacy considerations mean you shouldn’t use them for resume screening. However, using private, secure AI recruitment tools for candidate screening can significantly augment manual searching efforts.

Tools that allow you to search through passive candidate profiles and filter them based on your requirements have existed for years. But the advanced language skills of today’s AI tools makes them much better at creating accurate matches for candidates that may have slipped through the net.

Large language models (LLMs) understand natural language input, whether typed or spoken, which allows users of LLM-powered search and match solutions to enter queries in their natural language. This makes Boolean searches easier and gives experts another way to articulate their talent needs and return a higher quantity of relevant results.

Just make sure that AI is used to complement your manual talent search, and not replace it completely. Many developers still struggle with fixing algorithmic biases present in their AI tools, which can lead users to unintentionally discriminate against certain candidates.

However, this serious issue may be less of a problem in the future. For example, an LLM-powered matching solution built by a global financial services company outperformed 3 commercial (non-LLM) matching solutions in tests for hidden biases. Intelligently, the company removed names and other demographic information from applications before passing candidate data to the LLM.

A study from AI research company Anthropic showed similar results. By simply telling their LLM it is illegal to discriminate and instructing it to ignore demographic data, they came close to eliminating bias entirely.

categories

tags