In this blog, we explore the rise of chatbots in financial services, we discover who is talking about them, what they are saying and how financial services organizations can embrace this evolving technology.
Conversing with a machine forms a part of many people’s daily lives. From asking Siri the weather, to deeper philosophical questions such as ‘Alexa? What came first - the chicken or the egg?’, interacting with bots and automated voices is becoming more commonplace. The combination of AI, machine learning (ML) and natural language processing (NLP) has enabled organizations to build smarter chatbots in response to these growing consumer demands.
In financial services, both start-ups and traditional banks are looking to disrupt the industry by developing responsive, intelligent chatbots to transform their approach to customer service.
This got us thinking, what do people really think about chatbots? Using Brandwatch, a social media listening tool, we were able to begin analyzing conversation around chatbots online.
We gathered a years’ worth of data, from January to December 2019 to ensure that our analysis was an accurate representation of online opinions regarding chatbots. With over 12,000 mentions throughout the year, it was crucial that we drilled down into the details of this vast conversation.
We started with higher-level data, like key trending topics, that would allow us to narrow our search further. This is what we found:
As you can see, customers, experience, time, technology and business are just a handful of the most used words for this topic. These results were not surprising, so we looked again at keywords relating to our specific search query, to see whether the language being used was reflective of the overall sentiment towards chatbots.
This ‘word cloud’ above provided us with a little more insight. We could see that ‘customer’ appeared four times, alongside positive language such as enhance, automate, enable and improve – but this was only the beginning of our search.
We then explored demographics to see who was leading the conversation on this topic. The majority were male – and this is perhaps a reflection of the lack of females in tech roles today.
This graph shows the top professions of people discussing chatbots online. While the ‘artist’ category is curious, the other chatbot conversationalists’ job titles are largely what we would expect. More interesting however, was the actual sentiment of those conversations…
This graph shows that there is a significantly higher number of ‘anger’ related mentions, in comparison to ‘joy’ mentions (1,100 more from January 1st – December 31st 2019). When we investigate the language used in the ‘anger’ category, the overarching theme is that many people find chatbots less helpful than interacting with a human advisor.
Jibran Ahmed, Head of Research & Development at Capco UK believes that “in the interest of reducing the volume of queries coming to call centre staff, firms try to automate too much, and focus more on reducing their own costs and not enough on improving the customer experience”. It is important to remember that chatbots at this stage cannot resolve all queries. He added: “a chatbot is just one part of the customer journey, it must fit within the entire jigsaw puzzle that is the customer’s interaction with the organization’s services”.
We were also able to gauge the popularity of this topic by looking at the number of tweets, retweets and impressions for the top five keywords relating to chatbots. Since the beginning of 2019, the hashtag #chatbot has gained 11,137,640 impressions which is evidence of the relevancy of this topic.
To provide further insight, we did the same research for 2018 data, and this is what we found.
Interestingly, 2018 had the same top five hashtags as 2019, just in a different order. Despite an increase in 2019 of tweets using the ‘chatbot’ hashtag, overall, 2018 had a higher number of tweets, retweets and impressions.
Capco’s Jibran Ahmed believes that the large volume of mentions in 2018 coincides with an increase in chatbot implementation – that has now petered off. He said: “As we moved into 2019, chatbots became somewhat mainstream as a result of newer, more exciting technologies rising in popularity”.
There was continuous discussion on chatbots throughout 2019, with an average of 80 mentions per day. However, the larger, more dramatic spikes found in January, June and November have an average of around 220 mentions per day, which is over a 30 percent increase. These spikes tended to occur when a company announced a new chatbot service or technology.
So, what does the future of chatbots look like, and how can companies overcome the challenges commonly raised by consumers online? Here are Jibran Ahmed’s top tips for successful chatbot deployment:
1. Chatbots can’t resolve every query – firms need to consider what challenges a chatbot can help their consumers overcome. Start with requests that are simple but occur often - perfect those - before building more complex cases over time.
2. Consider the hand-off point to a human – when a bot can’t resolve a query, it should be forwarded on to a human who can assist. It is crucial that this process is well executed, seamless and efficient, to avoid frustrating the consumer.
3. Consumer education is key – customers need to adapt to using chatbot technology. Make the process simple and provide relevant instructions if needed. It’s also important to make the customer aware they are speaking to a bot, NLP is relatively primitive in comparison to human ability, so keep the customer aware of who they are interacting with.
4. Think laterally – The Cleo chatbot, for example, is designed around customer interaction. Rather than focusing on automating back-office processes, it focuses on building a customer-focused self-service money management product that is easy to use. Basic chatbot usage is now becoming a commodity, so firms need to look at building pro-active bots, who can execute a wider range of tasks for consumers.
Over the coming weeks we’ll be sharing the top online trends and topics of conversation in financial services today. If you want us to analyse a specific topic, email firstname.lastname@example.org.