This Week, In Recruiting - Issue 177
Sponsored by our friends Worksome
Worksome removes HR, Legal and Finance friction between companies and independent contractors at scale. Managing a significant number of independent contractors in your business? You need Worksome!
Connect here with a member of friendly team today.
Open Kitchen: AI & Automation is Eating Recruiting. Can We Find The Human Premium?
One of the soft themes of Brainfood Live (my Friday Livestream - you can watch it on my Youtube channel) over the past few months has been an exploration of the 'State of AI in TA'. We've had over a year of experimentation now and it was timely to see the very latest developments in GenAI-infused recruitment technology. This culminated in last week's live demo of about dozen tools which own ways sort to automate various parts of the recruiters activity - if you haven't yet watched the show, you can do this here.
We were not surprised, but surely impressed, by the tools on display. The obvious efficiency gains that these tools provide to the recruiter at desk smooths the way toward mainstream adoption. This is going to mean the recruiter role will be reconfigured by the tools we now have close to hand. And the search to find the human premium in the hiring process is now on.
I'm going to use today's Open Kitchen to discuss some elements of the recruiting funnel where AI is already making an impact.
1. Writing Job Descriptions / Job Adverts
Let me start by saying I fully appreciate the craft that goes into great copywriting. Experts like Mitch Sullivan and Katrina Kibben will continue to supply outstanding value in any era of AI. In fact, AI would do well to train their models on their work. That being said, AI generation of job adverts provides 'good enough' outcomes today for it to become the dominant way in which they will be created. It's like Nokia / Iphone transition - the switch will come as soon as the GenAI becomes infused into the job description edito of every ATS, all JD's will end up being produced this way
And, according to Workable's Nikos Moraititis, AI generated Job Adverts outperforms Human Written Job Adverts in terms of raw response rates. When you have a better and cheaper way of doing something, you're going to change the way you're doing it.
2. Structuring the Hiring Process
One of the most interesting aspects of AI is that the output can be used as input for the next stage in your process, creating a cascading series of innovations which speeds up the entire supply chain. How about using your AI generated job description as the seed data to generate a recommended hiring plan? An AI generated / human verified job description can be the starting point for the generation of a recommended interview plan, suggested interview questions, even recommended interviewers in your business, who have either proven interview capability, demonstrated affinity with this specific candidate population, capacity to do the interviews (because AI also has access to calendar...) and has previously interviewed and passed the best performing employees in the business. Both interview note taker BrightHire and AI interview scheduler GoodTime can generate plans like this.
The best thing about auto generated hiring process? It's highly persuasive to Hiring Managers and C-suite, because its not based on the quality of the relationship you have with these stakeholders and on the actual data on what has worked best in the past for your business, benched against what is working at large in the market.
3. Sourcing the Candidates
Sourcing candidates is the fundamental pillar of the recruiters job; companies hire specialist recruiters because they feel they cannot rely on job adverts to fill their hiring needs. Almost inevitably, sourcing was a core component of the capability of these specialist recruiters -n that is, the ability to acquire candidates who would not otherwise apply to job ads. The golden age of sourcing began in the late 00's - when recruiters realised that people were putting up information that could be useful for recruiting in their normal day usage of social media. 'Sourcing' are the art and science of being able to discover candidates by tracking this digital exhaust. And it became a valuable skill because this was not easy to do.
With the advent of GenAI, it has become easier to do. In fact, it may have even become trivial to do. The core capability of large language models is the processing of huge volumes of text and making some sense of it. Hence is should come as no surprise that GenAI sourcing tools - like hackajob intelligence - can instantly produce relevant shortlists of candidates in response to a simple natural language prompt. For those who are not customers of hackajob intelligence, which source only candidates on-platform, you can go to any number of providers which offer candidates scraped from the public web - PeopleGPT for instance. Also note how 'the big two' of candidate databases - LinkedIn and Indeed - are both changing their search functions. Rather than requiring skilled sourcers who have invested time and effort to learn complex boolean searches, they are both now moving toward natural language search to activate the AI to recommend the candidates.
As with Job Description generation, the results do not have to be perfect, they results just have to be good enough. And with the efficiency gains given back to the recruiter, they almost always going to be.
4. Reviewing CV's, Resumes, Applications
In 2024, recruiters have been overwhelmed by applicant flow. The global economy is in far worse shape than domestic commentators are willing to articulate. Inflation has increased pressure on workers to seek better pay, whilst that same inflation has increased cost of doing business, leading to layoffs, suppressed growth and lower job volumes. We've not been able to handle it, as the innumerable TikToks and LinkedIn posts from exasperated job seekers signifiies. When Kevin Grossman releases ERE CandE's report, I fully expect the reports of candidate experience to continue on the downward trajectory.
AI applicant review is here and, despite ambiguous legal standing (we're doing a Brainfood Live on implications of the EU AI Act - register here), overworked recruiters are going to welcome it. Like sourcing and job description generation, this is within the sweet spot of LLM's can do - review large volumes of text and sort order. We saw demo's of this from at least 3 of the vendors in Brainfood Live last week and I suspect this feature will become commodified quickly, with the key differentiators going to be UX, configurability and how the well the vendors stay on compliant to the spirit and letter of the legal jurisdictions they operate under.
The strongest argument will be factual - AI applicant review will be less bias than human review because it eliminates the variability that inevitably comes with multiple human reviewers. You ask 100 recruiters to screen the a longlist you and you will get 100 different shortlists.
5. Screening Calls
I thought we would be several years away from AI taking on the outbound element of recruitment work, especially on voice or video. Having now gone through process as a candidate on several different systems, got familiar with the CX, I can see it that is another area where AI will rapidly become default and replace the human action. In the demo, we saw 4 examples of this from Talkpush , Scotty, Braintrust and Round One. Whilst I screwed up the answers on most of these, we can be sure that real candidates who have both time and motivation to pass the screen would do it. I think the journey will be make will be similar to dealing with Chatbots on websites - there are circumstances where the candidate might prefer speaking to a bot rather than a human recruiter - less cognitive effort, less small talk, greater efficiency in processing, on-demand according to candidate schedule etc. Humans will still in recruitment, but not on the first screen, when the questions are fixed.
Again the voice bots are not perfect, but it's pretty clear that we are on a path of constant improvement. They will get better and I suspect that many are already good enough. As a caveat, I would say that bots on screening calls are at this moment best deployed for job advert applicants, where candidate motivation for the job is likely to be high. However, we can anticipate a situation where a bot outbound call to schedule an interview with a human recruiter to a sourced candidate would not be outrageous. This is coming and I'm here for it.
6. Interview Scheduling
How much time is still spent coordinating interviews?
I spoke to a ATS vendor last month who conducted research for one of their premium enterprise customers and they were able to reveal that schedule 1 x candidate interview on average took 37 minutes of time per candidate, a figure calculated by email composition, slack messages, telephone calls to all the stakeholder's in process to arrange a time. This is a conservative figure as it does not include the time invested by those other stakeholders. Now imagine this scaled out across hundreds of interviews - maybe thousands over the course the year - an incredible time synch.
Automated interview scheduling is likely to be one of those 'low hanging fruit' projects which any serious Head of TA will be looking to implement if they have not already done so. Plenty of great providers out there if you're looking for 3rd party solution - GoodTime, Cronofy and others will all do this. Recruitment Co-ordinators are already shifting to Recruitment Operations, expanding scope into other areas of recruiter enablement where recruiters can stand to benefit from more and better use of technology. Creating more time in the recruiting team, which can then be turned toward other ways to create more time - the efficiency wheel starts to spin.
7. Interview Note Taking
I've been banging on about interview intelligence for a while now. It may be the first time recruiters genuinely experienced AI 'at desk' and saw it immediately provide obvious value. The idea that you can simply eliminate your note taking and focus on the interview was unthinkable 5 years ago. Yet here it is today, as a normal tool which most recruiters should now be aware, if not already actively using. My enthusiasm for the category would have abated were it not for the fact that this simple solution to a simple problem is generating more innovation and greater value seemingly with feature release.
What happens if all your candidate interviews are recorded and transcribed? The AI generated transcripts themselves become starting points for further innovation which improves your organisations hiring capability; how about being able to resurface up candidates in the db when a new role comes up, by asking the system (in natural language, of course) to show you all the candidates who talked about X issue or Y technology over the past period? Interview intelligence increases the value of every candidate record, revives potentially dormant candidates, reducing the need to advertise or source, and - because that candidate has been interviewed before - reducing time to hire if they now become relevant to the new role.
8. HM ScoreCard Generation
Some of the most valuable applications of AI are going to be the least exciting 🤣. However, any recruiter who has struggled to get feedback and scoring data from Hiring Managers will fully appreciate the value of auto generating HM scorecards. No slate of HM's are going to be consistent in producing feedback and scoring, and one laggard here can slow down the hiring pipeline for the entire organisation. Providing a tool which can listen to the interview, conduct sentiment analysis on the key questions asked and generate a human editable scorecard is going to accelerate the process. Think about it from the HM POV - they are not the core user for HR systems, and therefore cannot be expected to be fluent with report generation. Providing simple tools that do the work and only ask for review / edit / confirm is what is required for non-specialist users of software.
9. Automated (Yet Personalised) Candidate Feedback
Closure is important in any activity where emotional investment is at play. Failure to provide timely or adequate feedback for candidates who have been assessed is the leading cause for candidate dissatisfaction. The generalised disclaimers of 'not being able to get back to every candidate' might be been adequate in 2014 but no longer fit for purpose in 2024. And yet the provision of feedback for every candidate is practically impossible - until the advent of GenAI.
There is no need to disguise that the responses are from AI. In fact, from early 2010's research, we understand that job candidates can often prefer dealing with a bot, especially when the purpose of the communication is specific and outcomes based. What candidates want is a responsiveness, a sense of fair treatment and useful feedback that can be used to improve their chances in the job market. Popp AI gave a great demonstration of this in Founders Focus a couple weeks ago - using data sources like the job description, candidate CV, interview transcription, interview scorecard, previously transcribed dialogue - to able to produce accurate and valuable feedback to candidate. This is going to become standard CX; we will human recruiters evacuated from this part of the process, not only as a recruitment efficiency, but also because of litigation risk.
10. AI Talent Relationship Management
We have seem some impressive products emerge in the Talent Relationship Management space over the past few years. These tools emerged when employers realised that you couldn't just sell top tier candidates 'on the job' but had to cultivate long term relationships with the communities you most want to hire from. However, pre-AI, these tools were all pretty much if-this-then-that messaging tools, which were decen but blunt and in no way really personalised, responsive and intelligent.
All this changes with AI Talent Relationship Management. Now we have tooling which can automatically track candidate behaviour, learn from their digital activities how motivations are changing and conduct hands-free re-engagement on a human recruiters behalf. As a former recruiter whose strength was in relationship building, tools of this type would've made me 10x more productive. And if, I were an agent, made my 10x more money.
How long before mainstream depends purely on the product market fit - the only thing keeping it from ubiquity is cost.
So Where is the Human Premium in Recruiting?
I'm stealing this term from the great Peter Hinssen, who talks in general industry about the impact AI will have on the world of work. I am confident the human premium will be found, especially in moments when the problem statement is undefined, the collaboration is necessarily intense and the way in which the service is delivered, is more important than the outcome of the delivery. I'll take more about this in Part Two next week, as obviously this essay has gone on too long.
Question for you before I go: where do YOU think recruiters will find the Human Premium? Comment below folks, worth starting a thread on this...
Now out of the kitchen, onto the lounge 👇
What's Going On?
Big List of Recruiter Events to Attend in 2024
Big List of Recruiting and HR Events to Attend in 2024 - updated folks. Coming back from Recfest, no question my enthusiasm for in-person events has only grown. Make sure you check out this spreadsheet - add any you think are missing - and get attending some of these amazing events. Btw, we might as well start adding 2025 events also, so get cracking with this also!
Brainfood Live On Air - Ep268 - How to Build a Job Scraper with Python (as a non-coder!), Friday 16th August, 2pm BST
This is probably the final AI / Automation focused show of the year. Please review the Youtube channel here for the free archive. This show is going to be the DIY version of AI ? Automation - you've seen the tools you can buy, but what about the tools you can make? Sourcing legend and former job seeker Aaron Lintz, Sourcing Specialist (Black & Veatch) did just that - building a web scraper which crawled the internet for Talent Acquisition jobs in real time. We're with Alla Pavlova, Recruiter (Riot Games), Ach Petrosyan, Senior Recruiter, Alexandra Gyvetvai, Recruiter & friends. Register here
Employer Story Video Tech for Talent Attraction, Tue 27th Aug, 10am BST
So video is something I personally need to get better at, so I am very keen to understand what is working when it comes to short form video for recruiting. VMJ have been a leading player in using video to promote jobs, but we know that there is so much more to the medium that job ads - what other forms of video content work? We're on Tues 27th August, 2024. Register here
Founders Focus - Ep50 - Up close and personal with Maxime Legardez Coquin, CEO of Maxi, Tues Sep 3rd, 2pm BST
We're back with Founders Focus, the interview series where we get up close and personal with the people changing the way we hire and work today. Next up, we're with Maxime Legardez Coquin, CEO of Maxi. What drives Maxime to become a serial entrepreneur? What can we learn from his entrepreneurial journey? How does the world look when the resume is no longer the key source of candidate data? All this and more - register here!
ERE Recruiting Conference, November 12-14, Anaheim, CA, USA
As I mentioned, I'm stealing Peter Hinssen's term, 'human premium'. I suspect that it, along with access to collective intelligence, is going to be the decisive component of human work in era of generative AI. How does this translate for us in TA? I'm going to be presenting a discussion on how recruiters can future proof in the era of AI. I would welcome any comments or contributions from the community in advance of this talk. See you in Anaheim: register here
NB: if you want a discount code, comment "ERE" below and I will DM you
If you have an event, webinar or podcast going on next week and want it featured on next week's newsletter, comment below with the link and event details. Don't forget to at mention me so that I see it
What Are You Doing?
Sophie Bullock is launching a new job search buddy service, Landed. Love this idea - too many of us are spending way too much time doing this hard task on our own. Very important to increase the density of your job search network - the perspective and intelligence you get from it will make the difference!
End Notes
Sun is still shining and I keep thinking this is the last week of summer. I have a feeling we're going to have a lot to deal with in Q3 and Q4, so I am going to this period as much as I can before everything gets too real.
Have a great week
Cheers
Hung
Hung Lee is the curator of Recruiting Brainfood, and now This Week In Recruiting. Subscribe to both if you are into recruiting or HR or just interested in world of work.
Talent Acquisition Strategist | Greater Seattle
3moThis is great info and shared tech that I wasn't aware of. Thank you!
Talent Acquisition
3moFantastic insights as always, Hung! It’s clear that we’re moving towards a landscape where efficiency and use of AI in recruitment are setting new standards, allowing recruiters to shift more towards strategic advisory.
AI is changing the world - I am here to supercharge that change | Connecting HR and Tech | 12+ Years Leading People & Product Initiatives | opinions expressed are my own
4moThis post made me recall the first episode of Brainfood live which was covering ChatGPT and how it can be used by recruiters. Fast forward today and every respectable ATS has integrated LLMs in their features. Writing Job descriptions with AI is the new normal.
High Performance Recruitment Coach - I help high performing recruiters become high performing recruitment leaders
4moThanks, Hung. This is exceptionally well thought out and explained. The human premium element will be fascinating to see—will we have fewer but much more valuable (hence highly paid) recruiters or more recruiters as AI-led productivity improvements simply generate new or additional related tasks?
Sales Development Specialist @ SocialTalent | - Helping growth enterprises Find | Hire | Onboard | Engage great talent
4moAI is definitely reshaping the recruitment landscape, but with so many tools available, the challenge now is training large teams to use them effectively. Leaders in recruitment are recognizing this as a critical step toward staying competitive. Your list has great insights into that problem well done.