Using AI to Transform the Banking Experience

At the FinovateSpring 2019 conference in San Francisco, AI was a hot topic. From financial institutions using AI to power everything from personalized customer experiences to underwriting, to fintech providers showcasing AI solutions, the technology was on everyone’s mind. In a panel titled “Which Artificial Intelligence Technologies Will Really Change Financial Services?”, we heard from experts who are harnessing this technology in an effort to exceed customer expectations and outpace their competition.

Beyond personalized customer experiences

Ankit Bhatt, Senior Vice President of Omnichannel Experience at US Bank, discussed how the organization’s goal is to become central to its customers’ lives. It’s not enough for US Bank to deliver personalized customer experiences — this is “table stakes” as Bhatt explained. Instead, US Bank wants to use AI to anticipate its customers’ needs and simplify their interactions with the bank so they can focus on running their businesses. This includes using AI to predict cash flows from small business customers’ transaction history so they can avoid overdrafts on their accounts. It also includes real-time account opening, loan approvals, and account funding so that business owners get access to the critical funds they need, when they need them. Bhatt explained that the culture at US Bank is focused on customer obsession, and that AI is not just about the technology; it’s about “bringing an experience to life.” Currently the bank is running close to 100 AI-based applications to support this goal. Bhatt and his colleagues at US Bank believe these experiences will develop fiercely loyal customers.

Where AI should be applied

When deciding how to use AI in a financial institution, the applications are potentially endless. Emil Matsakh, formerly Executive General Manager, Chief Analytics Office of the Commonwealth Bank of Australia, explains that machine learning algorithms within a financial institution should be focused on 3 core areas:

Improving the customer experience and their financial well-being through personalized insights
Optimizing risk-taking, vs. just managing risk
Enhancing productivity

According to Matsakh, Commonwealth Bank uses hundreds of machine learning models to support its AI program across 19 channels. The investment is paying off, as the bank has recently reported improvements in customer engagement of up to 400%.

Creating a culture of innovation

As an increasing number of financial institutions seek to adopt AI to create a competitive advantage, how can they truly embrace these new technologies? Jon Zanoff, Managing Director of TechStars, argues that for organizations to truly transform through technology, innovation should be tied to compensation. Employees need to be provided incentives to test new technologies, with an understanding that not all projects will be successful. Employees should be rewarded for taking chances and for working towards the goal of “doing something smart for customers,” as Zanoff explains.

What about the data?

With any AI strategy, sources and volumes of data must be considered so that machine learning models are trained and refreshed effectively. But where does this data come from? It’s a combination of pre-existing data plus real-time information that is collected as customers engage with the financial institution. However, organizations must develop strategies that encourage customers to provide more data, especially in light of the erosion of consumer trust in financial institutions that has occurred over the past decade. The panelists discussed how creating more value for customers will encourage them to provide more information. For example, if banks can demonstrate how customers’ information will be used to save them money vs. to sell them new products, this helps earn trust and creates an environment where customers are more likely to provide the data needed to power machine learning-based solutions.

Getting started

As financial institutions continue to adopt AI, the time to get started is now. Begin by identifying a clear objective: Where do you think AI can provide the most value for customers? Then develop a proof of concept with clear milestones and ensure you have a data strategy, so that your machine learning models are effectively trained. — At Appen, we’ve helped leaders in machine learning and AI scale their programs from proof of concept to production. Contact us to learn more.