Curiosity kills the (less efficient, more costly, and hopefully-not-but-possibly-biased) cat | Probe Group

We tried something new last year.

*waits for applause*

Okay, so hopefully everyone tried new things.  And to be fair, we did quite a LOT of new and interesting things – in automation, AI, analysis, augmentation – but there is one thing that we tried that turned out to be even more interesting than watching a neural-netted Mario try to level-up.

Note: I was going to write a blog specifically about how we structure our approach to AI, but instead I thought it’d be nice to share a story about finding the right problem, the right partner, and the right solution. 

This year, we decided to bring artificial intelligence into our already intelligent recruitment processes. As any organisation whose primary product is delivered through human productivity knows, recruitment is the first thing you need to get right. Bringing the right people into your business not only delivers the right outcomes for your stakeholders (good productivity, continuous improvement, high engagement levels); it also propagates your culture organically.

Recruitment can be an expensive cog in your machine, when you add all the testing and interviewing and administration time together.

What are our human recruiters great at?

Assessing capability in nuanced human scenarios, identifying culture fit, articulating and ensuring the employee experience.

What do our human recruiters find challenging?

Getting to candidates who are only available for screening after hours.

For all the wonderful candidates we speak to, we were only getting to 60% of the application pool.  So many of the people who want to work with us already have jobs or other daytime commitments that an awkward game of phone tag ensued. With our ability to screen from any location, we have a broad span of hours – but what about weekends? Public holidays? The dead of the night?

How could we get to more people, without needing more people?

We took a look at resume screening, but – and stay with me here – a resume matters much less to us than a conversation. We investigated video screening, and while the technology is cool, what didn’t feel right for us is that so much of our communication with customers is faceless. We are words on a screen; we are voices on the line; we are actions behind the scenes. 

So when we found a voice-only screening solution, available 24 hours a day, we were interested. When we found out that it analysed what candidates said for behavioural attributes and preferences, we knew we had to test it for ourselves.

Did it work for us? You bet it did. We set up our digital recruiter, Ella. Ella sounds engaging but digital; she doesn’t pretend to be a human, although we’ve given her a female voice and name so it’s not completely alien. Ella functions to the right level for screening – she doesn’t do small talk, she doesn’t go off on tangents, and she doesn’t react to people stumbling or saying something peculiar.  She has no bias about names, backgrounds, gender, age, accent or intonation.  

What is our digital recruiter great at?

Listening to what our candidates say in standard screening calls – rather than the way they say it; interpreting how people like to work.

What do we hope our digital recruiter will nail?

Predicting whether people are likely to stay and how we can best set them up for success.

What Ella captures gives us a great starting point to identify people who are likely to thrive in our open book, metric driven, highly structured work environments. We are correlating the attribute information returned through the natural language processing engines with our on the job performance to improve how we use the data.    

While the screening call is just the first step, this information helps us to prioritise the next interview stage, and tailor our human recruiter interviews to help tease out underlying preferences to ensure we’re setting people up for success when they join our business.  The process remains human-led, with AI providing us better data to make better decisions in a more efficient way.

That primary issue we faced around accessing candidates disappeared, with 100% of applicants given the opportunity to complete a screening call at a time that suited them.

31% of the candidates completing their screening with Ella are doing it after hours and on the weekend. 

There have been other benefits too. For the business, we’ve managed to reduce our screening costs by 65% just through incorporating this technology in our process.   And for our human recruiters, they are able to apply the attribute insight to our shortlisting to highlight better candidates. Our hire rate from the shortlisted candidates has jumped up 5%, and we’re still tuning the process.

This last bit is important. This innovation hasn’t been about elimination; it’s process augmentation that allows humans to focus their time wherever empathy, decision making and human connection matter most; while we let the robots (sorry, Ella) handle the repetitive and structured parts of the process.

What has the employee experience been like? 

“It was actually better speaking to a robot rather than a human as (there was) no emotion when you answered the questions.  Every question had a purpose.”

“Speaking with Ella was interesting and it made me feel comfortable about the future.”

“(Ella) would ask questions after you answered that related to the first question.”

Nice work, Ella.

We picked a great tool and honed the way we used it to maximise the value. That’s the trick with all of this AI – the tools are great, but the way you use them is what matters.

Contact us if you’d like to chat more about what we’ve learned.

It’s a pretty exciting and intimating tim

e for decision makers: there is so much new, interesting technology taking different approaches to solve business problems.  What can be intimidating is knowing where to start, so aside from sharing how a little thing is a big win, I wanted to give you some quick ideas on how we’re tackling that challenge.


Pick a Problem

We all have loads to choose from; some of them are big and complicated, some of them are white noise.  I try to pick one of each to tackle at a given time.

Why bother with the white noise?

Solving a simple, noisy problem will have a smaller business benefit, but the mental cost of thinking about it or addressing it is a wasted cost, and worth eliminating.  It also gets you started faster, and what you learn on those smaller problems helps to clarify or shape the way you tackle the bigger one.

Why kick off something big?

The design time on this one is longer, so I start to frame up the way we’d like to solve it quite a while before I plan to get to the implementation; it gives me time to find the solution, and to take learnings from other projects over time to adapt and refine my plan.

Be Curious

Set at least one half day aside each week to learn – through reading, getting demos, or jumping on the tools and seeing what you can achieve in a few hours. 


Business leaders are trying to work out where AI fits in their organisation – probably secretly hoping for a panacea – but haven’t managed to get started.  So if you HAVE started, even on something little, give yourself a high five.