Blog article
See all stories »

Fintech sector’s overreliance on chatbots is a mistake

A recent study has found that 73% of consumer banking interactions are now digital, as people and businesses increasingly choose direct banks over more traditional banking institutions. Some of the key reasons for this include a more frictionless experience, a lower defection rate, and crucially – higher customer satisfaction overall.

Having said that, “higher” is still miles away from “optimal” and given the edge fintechs have over incumbents, this should be cause for concern. As many of my peers have already noted, finance-related startups often downplay the importance of “human touch”, and try to solve every problem with (often overhyped) tech – chatbots being a paradigmatic example.

Chatbots on the march

The chatbot “revolution” kicked off in 2016 when Facebook made the Messenger bot platform the centrepiece of its F8 developer conference, with competing tech giants following suit with their own projects. By 2018, however, news coverage of chatbots had started to plateau once it became clear that early adopters jumped the gun, expecting too much, too soon.

In 2020, though, chatbots were back in the spotlight, riding a wave of reports about mid- to late-stage deals accounting for over half of all deals for the first time. This clearly indicates that by then, chatbots had reached a certain degree of technological maturity.

For instance, JP Morgan Chase reports that its chatbot software (COIN) reviews about 12,000 contracts per year, saving the company about 360,000 hours’ worth of labour. Successful applications have also been found in fields like health care, education, and many others.

Reasons for using chatbots include their ability to churn out immediate answers, manage simple customer requests, function 24/7 with zero days off, and save time better spent on more essential activities. Besides, people do actually prefer self-service over talking to someone when it comes to things like getting their credit card balance, submitting an insurance claim, or booking a vacation.

On the other hand, when it comes to more nitty-gritty financial operations, results are decidedly less impressive. Unsurprisingly, most people (and businesses) are rather cagey and sensitive when it comes to money, which puts a damper on the brave new world of chatbot-mediated finance.

Revolution bogged down

Fintechs these days serve many demographics, including the un(der)banked segment of the population and SMBs. To make this possible, instead of traditional customer support backed by an advisor or a customer service rep, they often rely on a self-service platform, which allows customers to make financial decisions with little to no human interaction.

As you may expect (or have already experienced yourself), this comes with a set of pretty serious problems that reduce the overall attractiveness of the fintech business model to customers.

“Sorry, I don’t understand”

Giving your chatbot a real name or a vaguely humanoid avatar will not conceal the fact that it’s not a real person. Since chatbots don’t understand natural language, faced with an unfamiliar request they either give no answer, leaving the customer feeling helpless, or lead them in endless circles. This is especially noticeable with rules-based chatbots triggered by pre-defined words or phrases that send them down a generic decision tree. More fancy, AI-driven bots that leverage NLP are slightly less irksome, but only slightly.

Lack of personal touch and emotion

Running into a financial snag can be really stressful. In situations like that, customers want not only help, but also a reassuring voice and some basic human empathy. Bots, however, typically lack access to customer data on the back end, making personalised service virtually impossible. This means that if the conversation fails to follow a specific path, the customer is in for a rude awakening, as the bot gets stuck in a loop of dull incomprehension. Besides, most people find artificial displays of emotion uncomfortable (or downright creepy) anyway.

High installation and maintenance costs

While cheap bots are relatively inexpensive, they’re also quite useless. For this reason, most fintechs go with a higher-end product. Apart from the initial cost of installation, however, more sophisticated chatbots require ongoing adjustment and optimisation to make sure they give the right information and represent the brand appropriately. As business priorities and user demand shifts, the bots need to be updated accordingly. This, in turn, requires a close analysis of previous bot conversations to identify common questions and replies. Leaving the bot to its own devices will give up the game almost immediately and lead to customer dissatisfaction.

Summing up

Chatbots are not unequivocally bad. In many cases, they work just fine as the first line of “defence”. But if you want your fintech to really stand out from the (rapidly growing) crowd, focusing on high-quality customer service is the ticket. A chatbot, no matter how advanced, won’t be able to help a customer whose funds got stuck in  transaction limbo or seemingly lost altogether. 

For a specific example, look no further than the infamous case of Robinhood. In 2020, tons of users complained about their accounts being looted and there being no way to reach the company by phone (24/7 customer support was then introduced in 2021). 

In short, even though most fintech customers are happy to do self-service for their day-to-day banking, the pressure value of a direct line of communication is crucial – fintechs are supposed to help “the little guy” and be more, rather than less personal than the big, hulking, bureaucratic incumbents.



Comments: (1)

Ketharaman Swaminathan
Ketharaman Swaminathan - GTM360 Marketing Solutions - Pune 30 April, 2022, 16:40Be the first to give this comment the thumbs up 0 likes

As I wrote in Can Chatbots Replace Humans?, I'm sick of human CSRs with room temperature IQ and prefer chatbots for many customer service interactions. But I do concede that, despite the passage of five years since I published that blog post, I haven't found too many examples of what I term "omnichannel customer service" with seamless handoff from chatbot to human CSRs.

Rafal Andzejevski

Rafal Andzejevski



Member since

07 Dec 2021



Blog posts


This post is from a series of posts in the group:

Banking Regulations

Discussion around current trends in regulations for banks globally

See all

Now hiring