A Look at ATS (Aplicant Tracking Software)

[00:00 - 00:12] So I want to talk a little bit about applicant tracking software. So applicant tracking software is a broader term to describe a comprehensive ecosystem of different functions that are used for hiring.

[00:13 - 00:25] And these tend to be either big platforms that integrate which your HR, which are external tools, whatever. They generally keep a database of candidates, job profiles.

[00:26 - 00:34] And some of them are a little bit more fancy. They also have tools for reaching out, like lead sourcing.

[00:35 - 00:43] So a lot of them will also offer data analytics on your recruitment, your hiring. And there's a lot of integrations with third party software.

[00:44 - 00:55] The thing to remember here is that these are not all built the same. A company might have several different types of applicant tracking software working together in an ecosystem.

[00:56 - 01:11] And it's also perhaps obvious, but most of these different types of software are integrating AI these days, right? And as we saw in the among of the previous slides, the biggest use for AI is screening candidates.

[01:12 - 01:20] So there's generally three classes of AI screening software. They're not mutually exclusive.

[01:21 - 01:28] Some applicant tracking software will do all three of these. Some of them will be focused more on just one.

[01:29 - 01:39] And there's other subcategories, but these are the three I wanted to point out. They work at different levels of abstraction.

[01:40 - 01:51] It gives you a good idea of top line top to bottom of the kind of scopes that AI and LML subscribers and screeners are being applied. So simplest ones are resume parsers.

[01:52 - 01:58] These actually predate LMLs, though they are using LMLs these days. I think it's most of them.

[01:59 - 02:04] They can just be algorithmic or rules based. So the concept is very simple.

[02:05 - 02:11] You have a job description, you have some keywords. I'm going to be referring to these as selection criteria.

[02:12 - 02:17] You have keywords that are selection criteria for a candidate. And you have a CV.

[02:18 - 02:34] You're basically doing a semantic matching between the job description and the CV. Right. So it'll pick up on things like if the job description needs someone that has Kubernetes, for example, and that's not on the CV, if you instantly flag that.

[02:35 - 02:42] Right. So this is a very kind of cost effective way to parse resumes.

[02:43 - 02:57] It's much cheaper than a complex LML architecture that generates some of the stuff we'll be looking at later. And yeah, an applicant tracking software is that have multiple different AI tools.

[02:58 - 03:07] This could be like the first filter. Now, AI summarizers and screeners are one step up from that.

[03:08 - 03:24] They will still do keyword matching unless the keyword matching the resume person is already happening earlier on in that process. These are generally LLM based. And they're trained on candidate profiles and job descriptions, which is a key thing.

[03:25 - 03:41] And they will look to summarize the CV and make some kind of judgment on whether the candidate matches the job description. Someone will generate a score or like a ranking.

[03:42 - 03:52] So someone that matches a lot of different criteria might score like 95% something. Someone that sounds completely unrelated to CV could get zero or five or something.

[03:53 - 04:08] And then these scores and rankings can in turn be used to select candidates or a driver shortlist. One step up from this, something that's a little bit more advanced is candidate enrichment.

[04:09 - 04:22] It effectively uses a lot of the saying tools as AI summarization. But what's the part is it produces a holistic candidate profile for CV.

[04:23 - 04:32] And it might produce it independently of a job description. So this is pretty useful for companies that are quite large and might have a lot of different jobs open.

[04:33 - 04:41] You get a lot of candidates across different roles. You build a list of candidate profiles with summaries, maybe notes about things that stand out about them.

[04:42 - 04:52] Again, you can still build them in a layer of ranking and scoring if you want. But you're creating something that's a little bit more enriched than just a simple summary and a score.

[04:53 - 05:06] These systems will often integrate one or both of the methods we talked about before. So there's a lot of ATS software brands out there.

[05:07 - 05:18] Cause for example, is resume parser. Freshworks is like a broader thing that has other streams, other admin stuff and different business functions.

[05:19 - 05:28] And it's a big picture, a enterprise tool that links them all together. I've only got five brands here, but this is a very competitive landscape with a lot of niches.

[05:29 - 05:38] So I looked at a couple of lists of top and hiring software platforms and AI resume parsers. You'll find a lot of different lists.

[05:39 - 05:46] They generally won't be in agreement with each other. The point being there are a lot of solutions out on the market right now.

[05:47 - 05:52] And a lot of them operate in different niches. Some of them are very specialized.

[05:53 - 06:09] There's some that have a really strong focus on integrating with things like ACORAC, another testing tool, some of them are really focused on lead generation. The point is it's not really possible to zoom in on any of them because there are so many of them and it is such a competitive landscape.

[06:10 - 06:26] Now, it's a competitive landscape also to get hired for a job, let alone to buy a screening tool for candidates. So I'm going to talk a little bit about the best practices, things you can do.

[06:27 - 06:33] This is mostly going to be theory heavy, right? But the best things you can do to avoid the AI filter.

[06:34 - 06:44] Now, if there's one thing to take away from this, it's that the fundamentals of a good CV sort of remain the same.