A number of organisations are either using or exploring using Chatbots for their customers, either as part of their website service, or through social media channels.


Dictionary result for chatbot:
a computer program designed to simulate conversation with human users, especially over the Internet.

Source: Google

We have had our eye on web chats (with human operators) and chatbots (with artificial intelligence) for a while, asking ourselves:

  • Is this the right thing to do for our public?
  • Will it make their lives better?
  • Will it assist with the demand pressures the police service finds itself under?

A day doesn’t seem to go by where you don’t hear something about AI, and there are some fascinating advancements and examples of this technology in use. As technologists ourselves, it’s hard not to get excited about what the future holds.

As predicted in Terminator, we’ll all eventually be wiped out by our evil computer overlords one day anyway so we may as well embrace it

As a digital team, our job is to use technology to solve problems for human beings. So what is driving organisations to use chatbots for their customers?

Trying to drive home a commercial sale

On some websites, if you linger on a page for a set period of time, a chatbot pops up “Hello, how can I help you?

Whilst the chatbot presents itself as your friendly assistant, the motive for its deployment is company driven, to try and nudge you to complete your purchase.

Assisting people who are lost on a website or to answer questions

In a similar vein, they are also used to assist people who may be lost on a website, unable to find what they are looking for. The bot attempts to suggest what they might be looking for and acts as an alternative to the site search.

They are also used as a kind of “frequently asked questions” service.

Replacing the one-to-one human conversation

Finally, they are used to replace the human conversation over the phone or in a web chat. This last method is seen as a cheaper, more efficient way to deliver customer service. After all, if the bots can have a conversation with customers then humans don’t need to, and employing humans is more expensive.

Again, this is driven by organisation motives, not customer focussed motives.

So what’s the problem?

What we think what is happening is that bots are seen as providing a solution to the following problems:

  1. Many websites offer a terrible user experience for meeting customer service needs – so the thinking is the bot will mask this issue and help the customer find things better.
  2. Call centre demand is very high – so the bot can assist the caller with their problem instead of a human, saving a phone call

Instead of fixing the root issues, adding a chatbot is putting a sticking plaster on top of your content, at the expense of the customer experience.

IT Pro did some research last year and found that 80% of customers don’t trust chatbots and would rather talk to a human. From those that used a chatbot only 16% were extremely satisfied, 39% somewhat satisfied, and 27% either somewhat unsatisfied or not at all satisfied.

Chatbot fail

Customer experience is much more important for the police

Whilst IT Pro’s research was conducted on users in healthcare and financial sectors, the trust rate is concerning. When our citizens need us they are often in a stressful situation and the last thing they want to do is wrestle with a chatbot. Whilst we don’t have the data, we expect the lack of trust levels would be even higher in our citizens’ case.

  • As a customer, have you ever used a Chatbot and thought, what great customer service experience that was?
  • As a customer, if you had the choice of using a Chatbot or speaking to a human, what would you choose?
  • When you are on the Internet and you want the answer to a question, do you search Google or try and find a Chatbot?

If the customer becomes frustrated by a bot, they will do one or more of the following:

  • a) type angry words at it
  • b) ignore it
  • c) leave the site go somewhere else

Police customers only have the choice of phoning the call centre or travelling to a police station if the digital offering frustrates them. A bad website not centred around user needs will drive up this call centre demand. A chatbot that equally frustrates will also drive up call centre demand instead of reducing it.

Another chatbot failure

Leave the bots, fix the actual problem and do it with the citizen in mind, not the organisation

Avon and Somerset Police have an award-winning high performing call centre. This is of course due to the professional and committed teams that answer the calls and run the centre.

We also have evidence to show this is also due to having an effective digital service that provides a viable alternative method of finding information and reporting online.

We aim that the service is at least equal to the experience of speaking to a call handler, if not better (and in some cases, we believe it is). 97% of all respondents surveyed who used our online reporting services were happy with the service.

We quoted our two favourite bits of feedback in our last blog post, but they are worth repeating:

“Thank you for a pleasant easy to use experience. I have Asperger’s and this made things a lot easier for me”

Upload video and images form

“One of the best online forms I have used”

Request a fingerprint appointment

We are not claiming to be perfect and this is an ongoing effort of continual review and improvement.

However, we aim to never offer a digital service that provides a worse experience than speaking to one of our staff.

Currently, we believe if we implemented even the smartest Chatbot technology in the world for our citizens, it would be a worse experience than they have now.

There is still a way to go for technology to get to the quality level we would need to avoid providing a sub-standard service to our citizens.

Don’t make me think

A powerful mantra from a usability perspective is that the user shouldn’t have to think about how to interface with a website or system, in fact with everyday objects.

One of the problems with chatbots is they actually hide the interface from the user, the same as a website search does. The user has to think about a range of possible options available to them. They don’t know what questions the chatbot has the capability of answering.

Of course, chatbots “learn” and become “smarter” over time with a wider range of interactions available to them, but the user doesn’t know that when they interact with it.

Increasingly chatbots are making recommendations and suggestions to try and counter this hidden nature of their interface.

A website that presents the options available in a clear interface doesn’t hide anything from the user, which doesn’t require them to think about it.

Again, our customers are often in a upset or distressed state, when humans are in this condition their thinking capability is reduced. This is why we can’t afford to make things any harder for them than they need to be.

Chatbots still have a place

However! We do see a place for bots in the workplace and in the not too distant future, we hope to begin some experiments internally for specific use cases where we feel they could add value to our employees.

This could be in the form of surfacing information and providing simple interactions, through to improving data quality and delivering process automation.

This is a very different user case from providing digital customer services to our citizens.


Here’s some further reading: