AI - Some practical uses in HR and recruitment

The second of two posts looking at AI (Artificial Intelligence). This one with particular reference to applications within Recruitment and HR
AI (Artificial Intelligence) is getting a lot of talk right now. This is the second of two posts looking at the subject - it's not a short topic. The first post looks at AI in general and questions whether it really does deserve all the hype it is getting in terms of; delivering useful services right now and in the future (AI is it over hyped?). This follow-up post looks at AI with particular reference to Recruitment and HR. If you work in the HR and recruitment space then the two posts will be broadly complementary. For those not in recruitment or HR, this second post should still prove interesting as it will provide a useful guide to practical applications of AI - many of which can be applied in other business areas as well.


AI for Recruitment and HR

As we outlined in our previous post round AI. It's is a topic that is getting considerable interest right now. The marketeers seem desperate to include as many of their products as possible under an AI tag. However what are the useful and realistic possible applications for AI in Recruitment and HR. We seek here to, strip aside the hype you may have seen, review what is usefully there, what works, and what - doesn't really work...

There are really only two reasons we can think of to look at using AI (or many other tools) and applying them to HR or Recruitment Activity:
1. To save time and by extension also save money and effort
2. Increase quality of service

With this in mind and after looking at quite a few market offerings with an AI dimension the following areas are the main ones that you can buy AI products and services in right now.


chatbots1 copy
A chatbot is where an AI system picks up emails or messenger enquires and seeks to answer them or provide some feedback. They can also be applied to answering phone queries. Done well, they can be a useful facility. They have some qualitative advantages over people - they are 24/7 - and they can provide a consistency and immediacy of service that also can improve quality. However - and here is the rub - from what we have seen, they need a lot of setup. In addition the economics re cost savings (if important to you) only really work if you can cut some HR staff after taking on the service. In essence, these systems provide an enhanced and more sophisticated form of FAQ (Frequently Asked Questions). This is all very well and sounds good. But, unless you do the work upfront and codify the queries and questions (and many of specific responses) then the service will look clunky and not be very useful. If you want to see a more general use of Chatbots just try the Google, Amazon, Apple services on your phone or Home Speaker device. You will quickly find that if the question/answer is not one that has been anticipated within the database then you will get a "I am sorry I don't understand your question". However if you have a big organisation they might save you some money or staff. And, if you can put the effort in they may also likely provide qualitative improvements in both HR staff services and also recruitment.

CV Screening

This is a modern enhancement of the CV parsing systems that largely fell out of favour a few years ago. The idea is that by using AI techniques the CV Screening system will be able to select those "ideal candidates" from the contents of their CV and put them to the top of your shortlist. The systems we have seen suggest they can screen out typically around 75% of your candidates up front. Sounds good. However if you use questions on your online application forms. You can, if you do it well, generally achieve about the same - and without any additional systems cost. Once you have sorted the Wheat from the Chaff there is no real substitute for a having a HR/Recruitment professional check over refined list of CV's and covering letters. Once the Chaff has gone this is typically very quick and by triaging your list you can refine it down very quickly to a shortlist.

The market will always have offerings that promise the power to predict peoples performance and behaviour. It's not necessarily false in every case either, AI can have a role to play here. However when we have looked at such systems they all required a good dataset of previous results with which the system can analyse and use to make its assessment. In this sense the AI system is playing the odds and numbers, by correlating past experience in making its choices. This is all very valid and computers are good at evidence based reasoning. However it does assume that you have the data (evidence) with which you can readily make such correlations, and have enough of it so that the system can be usefully configured. As we outlined in our first post on AI: we are largely still at the first of 3 stages, with current technology based round 'recognition intelligence' where algorithms recognise patterns. In practice, this seems to use to mean that using such tools productively means you need big datasets and lots of codified information on performance. So in basic terms you need to teach the system what good looks like. This is thus, an application area for the larger organisation that has both lots of in-depth historical information and in a form that is readily computer accessible. Beyond that we don't really see these tools - at least in their current form - of saving you any more money than what you can easily do with a good recruitment system. Form a staff or candidate perspective they don't really add to quality of service either, unless your processes are amended to also significantly speed things up.

Assessment Tools

Psychometric and Intelligence tools can be thought of as an early form of AI, in that they consist of models where a candidate or staff member inputs answers to questions and the system makes a determination round their likely attitudes, and sometimes aptitudes, based on a dataset. We are seeing these broader techniques (if not models) now being applied more generally especially across employees. Staff surveys and attitude measurement has come a long way. It is now possible to have much more sophisticated survey models and techniques that seek to tell you about what your staff is thinking and their attitudes. So, expect to see more of these tools with an AI label on them.

Beyond this, we think that video interviewing also offers good scope for more automation. We think the market will expand and that the costs of having video interviewing tools that can analyse a candidate's emotion, word choices and even aspects of their personality will become more generally available and cost effective. Interviewing is really a form of assessment. Assessing a candidate and making a determination as to whether they are right for your organisation is key to both external recruitment and the internal hiring process. Interviewing especially, takes time. A physical interview involves both candidate and hiring managers in a lot of effort. Video Interviews can save you time and be integrated into your recruitment process fairly easily. Combining them with AI tools that provide further assessment, seems to us, like a good use of resources. They may well save you both time as well as providing an increase in quality, given you can likely afford to put more candidates though the process. Note - we don't suggest Video interviews should always replace a person-to-person interview. You would likely not use it in every instance. However as a step in the process it seems a good use of technology and AI.

These are the main areas that we are seeing AI tools within HR and Recruitment. As we outlined in our first post (AI is it over hyped?). We think expectations round AI will retreat from sci-fi to a more mundane and more basic level. Indeed if we set aside sci-fi and look at where AI can work, and work well. We think that AI has a very promising future in augmenting human intelligence and activity with HR and Recruitment as well as many other business areas.

If you are keen to acquire AI tools, please do take care. This is a fast evolving market and offerings can change significantly over time. As our first post on this issue illustrated AI is much hyped right now. All that glitters will not be gold, and its likely many products and companies will very quickly come and go.