Generative AI has been a hot topic for a while now. On Episode 113 of the Talent & Growth podcast I spoke to HireSweet CEO Robin Choy about how we can use AI in recruitment and talent acquisition. We delved into how you can use programs like ChatGPT to streamline your processes and be better at your job.
How can generative AI be used in TA and HR?
TA has become a very text-driven job. We send a lot of debriefs, job ads and outreach messages every day. Everywhere a recruiter spends time writing text, AI can help. Outreach messages are one of the biggest ones for us, because people struggle with writing good outreach messages. There are best practices which generative AI can use to write a first draft. I say first draft, because you shouldn’t rely only on what’s generated, you should always revamp it and personalise it.
Job descriptions and job ads are another great use case. We’ve heard people using it to generate assessment questions for screening candidates. You put in a prompt like ‘I want to assess this skill. Can you list me a top 10 List of 10 questions that I can use?’, and that seems to be effective. If you need to show a candidate to a hiring manager, you’ll often write a quick blurb that can standardise the presentation. Generative AI can be fed raw data, then it’ll write standardised blurbs, which saves you time and helps to eliminate discrimination or unconscious bias because it levels the playing field for candidates down to key skills and experience.
How could generative AI make recruiters more efficient?
For a lot of the text we create, getting the information takes 20% of the time, while writing it up will take the other 80%. With generative AI, you have to fill it with the right information, because one of the rules is ‘garbage in, garbage out’. What AI does is allow recruiters to focus on getting that information and take 80% of the work off their plate.
If you write a job description for example, you need to put the compensation offered and skills required into the AI. As long as you’re feeding it accurate information, it’ll give you a good output and free you up to understand what the candidate will be doing during the first six months of the role instead of writing a job description. Learning how to gather information and write good prompts can save you 80% of the work.
You can also use Chat GPT to figure out what’s missing in a job description. Ask it ‘What type of information could I include to make it better? What’s missing in that candidate blurb? How can we make that candidate description more interesting for the client? What’s missing?’ It’ll guide you and help you find that information as well, which improves your output and conversions.
How can we get the most out of this tech as recruiters?
Always check its output because it’s often wrong. My advice is to be paranoid about it. I’d also recommend that you try to use it daily, because the more you use it, the better you’ll get at working with it. Try tools that fit your workflow as well. Tools like Jasper.ai for instance makes it very easy to write content articles. If you use a tool that rates job descriptions or job ads, it will be programmed with the best practices for that output. That tool will be able to measure conversions and do all these other things because it’s specialised. These tools add a layer on top of generative AI which can save you a lot of time. It’s become very important to be up to date on these technologies if you want to keep up.
How can recruiters use generative AI for outreach and messaging?
You can use AI to generate or review an outreach message that you’ve written. With reviewing, there are a few questions that you can ask. I’ll say, ‘Here’s an outreach message, what are the most fluffy, useless parts?’ and it replies with the parts of the message that are redundant to similar engineering positions. Basically it gives you better wording. It’s very easy to pinpoint and emphasise which parts don’t add value.
Another helpful function is using AI to improve the message and make sure it’s not biassed against minorities. For instance, you’ll say, ‘Here’s an outreach message, can you pinpoint which parts are likely to be biassed against a minority?’ and it will look for gender specific language and things like that which you should rephrase. Not everything will be right, because AI has a tendency to lie because it’s not actually backed with data, but it can be useful for making you think more deeply about certain phrases.
If you give a lot of context, you can get it to write a refusal email for a candidate. Give it directions like ‘Make it empathetic, explain what worked and what didn’t so that the candidate understands’. It can help you with candidate nurturing as well by keeping people up to date on their process. A lot of people don’t need to hire as much as they did a year ago, so we’re all thinking about how we can nurture our talent pool with outreach messages, email newsletters etc. to keep people engaged. It’s always a bit painful to get started writing a message, so go and ask AI ‘What should I say? How can I add value to the candidates?’ and it will come back with things you might not have thought of before.
Can people outside of TA use generative AI for recruitment?
If I’m a hiring manager and I need to add someone to my team, you can use AI as a recruiter by asking it for 5 to 10 questions to help you refine the role. Chat GPT will ask questions about the role so you don’t even have to input data, and then write a pretty accurate job description. Then you can feed it prompts like ‘Rewrite the job description to make it 30% shorter while keeping most of the information’ etc.
You can also say ‘I need to hire for this role. Here’s my notes, can you write a job description about it?’ The more context you give the better, so add context to make it very specific, which is more likely to drive a tonne of applicants. Just be straight to the point in a style that flows.
To hear more about using generative AI in recruitment, tune into the Talent & Growth podcast here.