Site navigation

Comment: Customers Don’t Want Chatbots, and Here’s Why

Martin Thorn


Martin Thorn, senior manager: Decisioning, Analytics & AI at CYBG, discusses cumbersome chatbot experiences and why people value human interaction over supposed ‘ convenience’.

I had to call my insurance company this morning. I tried to self-serve but it wasn’t possible on their website; mildly irritating but their contact centre opening hours were pretty convenient.

So, I picked up the phone, dialled the number and within two minutes I was through to a great agent who was able to solve my problem.

I am a happy customer, 10/10.

But surely it would have been better with a chatbot? That’s the future, isn’t it? Everybody seems to want them, don’t they?

Not a week goes by without someone trying to tell me that chatbots are the answer to a myriad of problems. However, it is always interesting when I ask the question: “When was the last time you had a good chatbot experience?

The answer is, almost, inevitable. A more interesting question, I find, is: “Would you choose a chatbot over a person?

Here is the thing; customers do not want chatbots. Customers just want their issue resolved. Now, that doesn’t mean to say that chatbots are a bad thing, it’s just that most people haven’t had their issue fixed by one before. If you need to solve an issue, you will automatically use the medium that gives the best trade-off between getting the issue fixed and being a pain – so most people call first.

It’s not as effective as swinging your broken toastie machine around your head in Curry’s, but it’s a lot less hassle. And those machines are heavy.

If the phone has always got a result in the past, then it’s going to the “go-to” for a long time to come. If customer’s don’t want them, why all the hype? Why are companies still pushing them?

It’s quite simple. On paper, chatbots will save companies a fortune. Ongoing costs are clearly much less than hiring a load of contact centre agents but there are a lot of hidden costs. The biggest issue, though, is that chatbots need data.

It all sounds rather obvious, but you would be amazed at how many conversations are derailed by the comment “so what data can we use to power the chatbot?” or “what questions is the bot going to answer?

Occasionally, you may find a gorgeous, well-maintained FAQ site, which would be a great start. However, what do you do if the FAQ site fails to answer or resolve the customer’s problem?

You either give a horrid “please call this number” message or you drop out to a human on a chat platform. If you do the former, then you are pretty much signing the death warrant for your chat engine as no-one will come back. But, if you do the latter, then you’ve left the door open at least.

Digit Leader Banner

It is often the overlooked downside of the chatbot – you also need a chat platform with a person at the backend. So why not start there? Buy a platform, train your agents and start your learning adventure. On day one, you won’t save a fortune, but as every month passes you will get better and better and, before long, your handpicked team of agents will be answering multiple customers at the same time.

In a past life, I frequently saw the best agents dancing from screen to screen helping up to five customers at the same time. It was incredible to watch and the customers were delighted. The agents were delighted too. Going back onto the phones wasn’t an option for either party- this was the future.

I went into the chat experience as a complete cynic, but after 12-months of working on the project, I became a convert. It’s not for every customer journey or even every customer but used int he right way it certainly is a fantastic medium. Once you have spent a few months or even a year you can start to introduce a chatbot at the start to help streamline or direct the chats and then you’re on the journey.

So, repeat after me – chat first, then bot. Maybe. If it’s appropriate.

  • Martin Thorn will be presenting at Digit‘s Intelligent Automation conference on the 20th of June 2019 at Our Dynamic Earth, Edinburgh. For details, visit

Martin Thorn

Martin Thorn

Senior manager: Decisioning, Analytics & AI, CYBG

Latest News

%d bloggers like this: