Improving Customer Experience using Chat

Chat can be a powerful tool for helping customers, increasing the service you are able to provide while maintaining scalability through deflecting easy cases using AI or curated answers.

The reason we wanted to move towards using a chat first tool was that we wanted to achieve several things through an in-app support experience:

• Improve the customer experience being available where they are working tailored to their needs and their specific plan/products (previously our support was on an external help centre only)

• Reduce the number of direct emails we receive by making the above a better experience.

• Implement human chat for topics where synchronous conversation will lead to a faster and more efficient resolution while maintaining tickets via email for topics which are more complex or technical

• Initiate more proactive support interactions when users are running into errors.

The process began with a vendor comparison, looking at our requirements against tools in the market and against the offering from our incumbent tool Zendesk.

Once the tool was selected the next step was to begin working on a rollout plan. At first this consisted of myself implementing all entry and exit flows as well as initial conversation flows. An important aspect of our approach was to roll out the experience iteratively so we could learn from each release and make improvements. Luckily this was made easier by the swarm model of our support team. 

Once the first team was rolled out brought in the team for the future iterations, being able to collaborate with my support colleagues was essential so I was not the bottleneck in our rollout. For this I developed a number of different documents from ToV, getting started guides, conversation tips as well as leveraging the vendors own training. This allowed the team to collaborate on creating initial flows for their rollout as well as develop processes and monitoring the tool looking for improvement opportunities.

In parallel during this phase a number of different workstreams kicked off. Excited by the opportunities a chat tool could offer our marketing teams responsible for our website reached out about using it to create answers to common questions on our pricing and blog pages. I was able to collaborate with these teams to curate that experience which eventually led to it becoming a strong channel for our sales team who started taking live chats directly from our website. 

As we looked to improve the experience further I was able to use my knowledge of Zendesk's API's to implement features whereby customers could add updates to already open tickets. We were able to experiment using our Algolia integration to serve better article suggestions for questions not covered by our flows and implement callouts to active incidents right in the chat so customers know we are on the case as well as preventing those customers from needing to reach out!

Through this initiative which was solely my responsibility we were able to go from 0 to 50% of our inbound volume being served by our chatbot over the course of 9 months with an average containment rate of 45% for conversations in that channel with no drop in CSAT.

Below are some images of some of the more advanced features we were able to bring into the experience.