Exactly a year ago we launched our chatbot – an intelligent, cyber-crime fighting chatbot focused on identifying an individual’s cyber security issue and connected them with help. This unique chatbot is a key component in our mission to ensure everyone in the UK had free, expert access to cyber security advice when they need it.

We take a look back at the chatbot’s first year and look forward to what we hope will come.

What does the chatbot do?

The chatbot is the front door of our services. When somebody needs help they come to our website and start a conversation with the chatbot. The bot then gets them to explain their issue – and using artificial intelligence – maps the users issue against our attacks database and makes a judgement on what they key issues are.  

It then does one of two things: 1) It connects the user with written web guides that provide step by step instructions for recovery or 2) It sets up a case with one of our volunteers who then work directly with the user to help them resolve their issue.

It works 24/7 and ensures everyone has immediate access to help to deal with the initial stages of the issue.

You can access the chatbot on our Get Help page.

How are you celebrating?

As a fully digital not-for-profit it is rare that we manage to get our volunteers together. To celebrate the chatbot’s 1st birthday we are throwing a birthday party in central London and we expect over half of our 30 volunteers to attend. There will be cake.

A bit thanks to TechTarget for hosting us in their London office.

Why did you build a chatbot?

The scale of cybercrime in the UK is huge – hundreds of thousands of victims a month. On the other hand, the number of cyber security experts is small – and the number who can volunteer their time is even smaller.

To meet the challenge of giving everyone in the UK access to free, expert cyber security advice we realised early that we needed to harness technology to help us.

By using chatbot technology we also realised we could operate 24/7, let users remain anonymous, provide consistent advice and keep our costs as low as possible.

How has it performed?

Amazingly!

We look at the chatbot performance in a number of ways including diagnosis accuracy, user satisfaction scores and volume.

They chatbot has processed well over 500 real cases in its first year (and hundreds of test cases) with a diagnosis accuracy over 85%. A much higher accuracy rate that we had forecast. Most pleasingly more than two thirds of our users scored their chatbot experience either four of five out of five for satisfaction.  

How will the chatbot change in future years?

The chatbot is currently part of a police pilot, with the Yorkshire & Humber Regional Organised Crime Unit being the focus region. We expect the chatbot will eventually work more closely with law enforcement and the reporting of cyber-crimes. We have already started tracking police reporting, to get a close feel for what types of cyber-crime are reported to the police and the % of crimes that are never reported.

As the volume of our cases increases we also anticipate that the chatbot will also be able to act as an early warning system. As soon as we spot new types of attacks we would be able to warn law enforcement agencies and the general public providing pre-emptive advice.

Finally, I think we will see the chatbot take on more and more of the role of guiding the users through the early stages of the attack response. There will always be a role for human experts, but allowing the chatbot to work deeper the response will allow the user to recover faster and put less pressure on our human team.

Happy birthday chatbot!