Can financial services chatbots fulfill their potential? It all depends on data

Photo credit: 3151940.

Bots are going ballistic. Transparency Market Research predicts the global chatbot market to be worth nearly US$1 billion by 2024. And according to bot specialists Personetics, there will be a surge in chatbot companies considering entry to the conversational financial bot space in the next 12 months, with over three quarters of surveyed financial institutions viewing chatbots as a near-future commercial solution.

But how can chatbots become effective? How will they engage audiences and transform personal financial management (PFM) into intuitive personalized digital assistance?

The appeal of messaging apps

Jake Tyler, CEO of Finnai, develops white-label personal banking chatbots for financial institutions. He argues that bots have found an audience by serving the immediacy demanded by today’s customer.

“If banks want to attract Millennials they will need to be where they are, on instant messaging platforms. Equally as important, chatbots are a way for banks to communicate with this generation in a way they are familiar with, by texting.”

Millennial tastes explain Facebook’s bot appeal. Facebook Messenger has over a billion monthly active users and more than 30,000 chatbots. Facebook beneficiaries include MasterCard, which will use artificial intelligence to communicate with customers through text messaging and speech, enabling account holders to check accounts, track spending, and review purchases.

Therefore, bots must be embedded into messaging platforms to speak to their customers. But how can they understand every end user’s financial position and make predictive assessments?

Categorization and aggregation technology

A crucial underlying data source for financial bots is based on account aggregation technology. This allows the personal finance chatbot to access all of an end user’s financial accounts and return a comprehensive picture of his or her finances. However, not every bot uses this. The best bots are only as good as the aggregation technology supporting them.

Without data aggregation or the ability to collate different elements of a user’s financial footprint, a chatbot will only interrogate one data set. Similarly, the limits of PFM were exposed when legacy banks were reluctant to share their customers’ data with each other, only allowing PFM tools to operate on their own internal data set. Using multiple bank connectors, fintech companies are able to aggregate from multiple banks in multiple countries, securely collecting a plethora of data sources.

But not every technology provider has this capability. Without aggregation technology, chatbots will fall short in supplying the intuitive, flexible personal assistant service that’s now demanded by end users.

In addition, chatbots will fail to deliver accurate results and advice to customers without quality categorization technology converting raw transactional data into actionable and meaningful data. This creates the personalized end-user data which the bot depends on. This technology relies on users’ reports of inaccuracies in the transaction data. The reports will allow the AI to learn from its mistakes, creating the personalized data for intelligent personal assistance and enabling predictive behavioral services for financial institutions and fintech manufacturers.

Chatbots in financial services do have tremendous potential, but that potential will be dictated  by the data they are built upon. Categorization and aggregation technology must be fully supported to give bots the foundation they need to personally benefit each and every customer.

This is an opinion piece.