Centrepoint

Building an AI Chatbot to help support young homeless people
Centrepoint

Share this post

Centrepoint is the UK’s leading youth homelessness charity with the mission to end youth homelessness by 2037. Since the recent COVID-19 pandemic, and the current cost-of- living crisis, Centrepoint have seen record numbers of young people contact their helpline. At times, the helpline staff cannot cope with the demand.

Centrepoint wanted to develop a chatbot solution to handle the increased demand to their helpline services by introducing a chatbot as part of their digital offering.

The chatbot’s goal was to enable Centrepoint to have the ability to interact with young people in need at any time of the day or night.

Implementation

Once we had been selected to undertake the project, we began by identifying the key use cases and designing the interactions users would have with the chatbot. With the help of the Centrepoint team, we developed the conversation flows that guide users to helpful information and points of contact who could assist them in their circumstances. 

Initially the chatbot was designed to only be available out-of-hours on Centrepoint’s website. The chatbot would then guide users to follow distinct pathways to access relevant information using button based conversation flows. The reason for this was so that Centrepoint could directly control the information young people would access without the risk of the chatbot providing the incorrect information. 

A key part of this project was to integrate the chatbot with Centrepoint’s knowledge base to look up information, such as local council information. When a user gives their location, we developed a function that would return the correct information for that particular area. This information was displayed to the user in a clearly formatted response, noting the key contact details etc…. 

The chatbot was built on our chatbot content management software, EBM. By developing the chatbot on EBM, we were able to add in the required functions to fetch data from Centrepoint’s knowledge base. On top of this, EBM allows the Centrepoint team to monitor and maintain the chatbot on a daily basis so that they can ensure the chatbot is performing correctly or add new content. 


Before launching, we shared the chatbot with Centrepoint’s wider team to gather user feedback on the chatbot. We gathered feedback on the chatbot’s icon, tone of voice, helpfulness and functionality. The chatbot was launched on the Centrepoint website in January 2024. 

 

Results

In the first six months since launching, the chatbot has received over 2,700 conversations and over 23,000 messages from web users. As well as being able to provide out of hours assistance, the chatbot has enabled Centrepoint to gather insights into the types of help their service users are requesting out of hours and how these differ from requests during work hours.


The future

The Centrepoint team were trained on how to monitor and improve the chatbot via EBM, and have since taken over these tasks. Now that conversational AI has been introduced into the charity’s digital strategy, Centrepoint are now looking at ways that the chatbot could be opened up to accept all user queries. This would mean allowing free text questions from users. This will give them even more valuable analytical data and understand how best to respond to young people in need. To achieve this at scale, we will be exploring using Generative AI and other technologies to provide answers to their user’s questions.

More to explore

Ready to kickstart your chatbot journey?