Bethan Blakeley: The art of automation

x Sponsored content on Research Live and in Impact magazine is editorially independent.
Find out more about advertising and sponsorship.

It’s something that has been around for a while and isn’t going anywhere – and yet it always manages to divide the room: automation. As with all things, there’s clearly a huge scale, with some extremes. A few are terrified that robots will shortly take over the world, and fear for our jobs. Some don’t believe in automation; it’s the root of all evil, and everything needs to be done “the proper way, like in the good old days”. Others feel automation is incredible, and use it at every possible moment, claiming it makes their lives easier, and makes boring processes faster and, well, automatic.

Although I’m frequently told I’m “not one to sit on the fence”, there’s definitely room for some middle ground here. Let’s get one thing straight from the outset: the robots are not going to take over the world any time soon. I’ve written, spoken and ranted many a time about how ‘artificial intelligence (AI)’ is neither artificial nor intelligent. As my childhood friend keeps telling people: “As long as the self-checkout at Tesco can’t work out when I’ve put my shopping in the bagging area, we have nothing to worry about.”

I get on well with my friend – but I don’t know if I wholeheartedly agree with her here. I don’t think we have nothing to worry about.

I’m not talking about artificial world domination. I’m talking about the small things; the things that seem unimportant and, yet, can make a big difference, particularly in our industry. The market research world is fast-paced, ever-changing and, sometimes, just feels a bit non-stop. Automation can help us navigate that.

There are small examples of automation in our work lives that you may not even recognise because they are so entrenched – computer documents automatically saving or Microsoft Office correcting your grammar. Others are fewer and further between, such as your conjoint simulators, your automatic text coders, or your social media scrapers. I’m not, by any means, saying the second group of applications is the ‘bad’ or the ‘ugly’ of automation. What I am saying is that it has the potential to turn bad or ugly – and it’s our job to stop that from happening.

Let’s step away from market research, just for a second. Pilots flying on autopilot or cars driving on cruise control are both great inventions when used correctly – but pilots aren’t allowed to leave the cockpit when the plane is flying itself. There is always a need for human interaction or, at the very least, human supervision or understanding when it comes to automation.

Here’s another anecdote: a friend of mine works in a hospital pharmacy, and they use a robot to dispense medications, among other things. This robot is several times faster than a human, much more accurate, and so on. Except that, recently, it broke. Staff weren’t able to override it, they weren’t able to log in and find out which drugs had been dispensed and which hadn’t, and they weren’t able to look at the patient records to do it themselves.

Panic ensued and the hospital ended up being put on a ‘code red’ (I don’t know the details, but it didn’t sound good). Staff had to re-do a lot of the records on paper and revert to the old method of DIY – a huge, expensive, painstaking process that took a lot of time and could have put lives at risk.

Moving on from near-death experiences and back to research, my point is this: automation is there to help you, not to do the work for you. There’s little point in having a ‘make segmentation’ button if you don’t understand what it does, what it needs, how it runs, or what it will give you in return. In fact, it’s not only unhelpful, but it could be damaging. I wince when I see yet another program touting how anyone can ‘run your own analytics and your own projects without the skills or the detail needed, in half the time’.

It might just be me, but I don’t think that’s the way we should be going. I wouldn’t want to hand over my data to a computer, and then stand up in front of my clients and present what the computer has told me, with no questions asked.

It’s us – the humans – that add the expertise, the knowledge, the context, and all the other good stuff. That is why clients pay us. By all means, get the computer to do the boring bits of your job so you don’t have to – but understand what it’s doing, check in with it, ensure that you’re happy with the outputs, and make time to understand them.

Like I said, autopilot is brilliant – but I’d still be a bit nervy if the pilot came and sat next to me for a nap mid-flight.

This article was first published in the October 2022 issue of Impact.