The 3 Biggest AI Opportunities in Financial Services in 2018

AI in the absence of a clean business strategy is just AI,” said Jeff McMillan, managing director, chief analytics and data officer of Morgan Stanley Wealth Management, at the AI Summit in New York. He said the goal is not to simply “do AI” — it’s to build an intelligent organization that understands there are certain problems it needs to solve by adopting AI as a tool, rather than a solution.

The current AI climate differs a bit from previous technological waves of change in several ways. The hype has become more feverish than ever before, the technology has become more complicated, and both factors are leading teams to spend a disproportionate amount of time talking about potential solutions and ignoring business challenges.

 Screen Shot 2018-03-07 at 12.32.29 PM.png 

The maxim that a problem well-framed is half solved is more true than ever.

The most successful enterprises will focus on identifying, framing, and solving today’s pressing pains and needs through AI technologies — not running experiments in the lab.

Three critical areas of focus are intelligent experiences (e.g., search and conversational experiences), cognitive insights (e.g., delivering more predictive and insightful information to inform decision-making), and intelligent automation (e.g., digitizing and automating operations with the aim of creating more efficient operations and more compelling customer experiences).

Intelligent experiences: the rise of the cognitive moment

This AI era is ushering in a new generation of digital customer experiences.

Some fintech startups, such as Change, use machine learning (ML) to predict when linked checking accounts are at risk of overdrafting. Uber uses ML to better predict routes and shorten wait times. Alexa is getting smarter about the information and value she can offer.

When a digital experience is able to anticipate our needs or offer incremental benefits without our asking, we experience a “cognitive moment.”

What is interesting about cognitive moments is that often, on the surface, it can appear as if little has changed: the user interface is not vastly different, the massive troves of data and technology behind the scenes are invisible to the user, and brands are not necessarily announcing, “Hey! We’re using AI here.”

With the evolution of voice, search, machine and deep learning, more responsive UI, and other intelligent technologies, there are countless opportunities for financial services firms to purposefully craft cognitive moments. Examples include the following:

  • Wealth management firms can adjust portfolio charts to display the information the client cares about, based on historical data, which means better retention and a more valuable and customized customer experience.
  • Brokerages can deliver an intelligent search experience that understands a client’s portfolio and goals and gives clients the ability to discover relevant investment opportunities.
  • Commercial banks can enable financial managers to save time by searching for specific data using voice rather than navigating spreadsheets; managers can perform eligibility checks or underwriting at the moment of data capture, enabling instant decisions.
  • Retail banks can reduce customers’ churning through tailored, personal insights or recommend a loan because a customer’s bank account is likely to overdraft in two weeks.
  • Credit card companies can automate the review of credit applications to quickly approve loans with less manual review.
  • Credit firms can also offer their members additional cards if they detect a child has turned 18 to cover things like college application fees or a detected change in life status (e.g., a last name change or marriage, in which a client would need an additional card for the new spouse).

For the majority of enterprises, the biggest challenge when it comes to AI will not be the technology, but the collective organization needed to become even more empathetic with customers’ and clients’ needs today. The best intelligent experiences will be crafted by emotionally intelligent businesses and built on mature data strategies.

Cognitive insights: more predictive information at the point of decision making

An organization is fundamentally only as strong as the collective decision-making ability of its individuals — and the decisions made from the corner office to the customer service frontlines.

AI will play a critical role in delivering more predictive data and information at critical decision-making moments. We refer to these sorts of deliveries as cognitive insights.

Cognitive insights are born from the middle ground of the enterprise tech stack, between front-end and back-office systems — think contact center software, CRM and email systems, customer and client account management software, and business rules engines. AI will increasingly make its way into this layer by resolving business problems before they get to the individual.

A meta-use case in this realm will see the deployment of AI to ensure other robots and algorithms are making the “right” decisions by studying the details, inputs, and ultimately the logic of how these systems arrived at their recommendations and decisions.

More commonly though, cognitive insight solutions will provide the logic required to make sense of the mounds of data, human knowledge, and intelligent predictions that we’re capable of gathering today.

Take, for example, the challenge of predicting which customers will become inactive and taking steps to deter customer churn.

Unsupervised ML can be leveraged to detect unintuitive correlations within the data from customers who have become inactive in the past and then use these insights to better predict which customers are likely to become inactive in the near future. Through this process, cognitive insights may start to predict that customers making frequent cash advances are more likely to churn, or that when someone stops paying their phone bill from their bank account they are more likely to abandon their banking services in an upcoming, pre-determined window.

More advanced cognitive insights systems will take the next step in proactively reaching out to the customers it detects are likely to become delinquent, based on leading indicators. Throughout this process, each user’s resulting behavior further trains the system to determine the right reach-out details, optimizing probability of conversion over time.

Screen Shot 2018-03-07 at 12.33.20 PM.png

Financial companies can capitalize on these solutions to improve time-to-revenue, decrease the time required to make a decision, and move expert knowledge into more central locations of the business. Examples include the following possibilities:

  • Wealth management companies can predict which families are likely to switch advisors when wealth moves between generations, increasing retention by sending customized messages to at-risk customers.
  • Brokers can predict which pieces of stock news and data the client will be most interested in and prioritize those items in client outreach to deliver better value.
  • Commercial banks can use customer insights derived from performance data and usage analytics to drive revenue growth by determining which of their clients will benefit from new service offerings.
  • Retail banks can evaluate personalized risk tolerance levels for clients’ portfolios through smart surveys then leverage that data to build accurate risk profiles on their customers. Banks can also predict which actions their customers want to take when they call the customer call center, open the bank’s website or app, or chat with a bank representative, making it easier and faster to retrieve that answer or perform that action.
  • Credit card companies can recommend the right card based on previous spending history and even Facebook or Twitter interests (if the user is willing to share that information).

Imagine the competitive advantage a firm would have if every employee had a team of analysts to call on day or night just before making a critical decision. The idea of cognitive insights is making this scenario a reality.

Intelligent automation: decreased touch points, increased throughput

One of the most commonly perceived applications of AI is its ability to create efficiencies throughout an organization by eliminating the repetitive, manual tasks that can come to consume entire departments.

Key legacy systems in the crosshairs of these advancements include risk engines, document management workflows, and employment management systems. Even on a small scale, these efforts can have a massive impact.

As a focused use case at Solstice, we teamed our AI assistant, Siena, up with our firm-wide Slack instance to solve a nagging space capacity issue. Ahead of all meetings, Siena now Slacks organizers and asks if they plan to keep the room. If not, it’s released back into the general pool. After only a week, meeting room capacity increased by 25 percent, with little to no negative impact or required behavioral change required from individuals.

It’s important to note the potential creativity made possible by these efforts. Intelligent automation will not merely be put to work reevaluating existing processes to more effectively automate the mundane, but it will also help firms re-envision entire processes. This is being done by combining traditional and long-standing robotic process automation efforts with process reengineering and AI technologies.

By implementing solutions within this domain, financial firms can rededicate their resources to the higher impact, high-complexity issues affecting their customers and their business in an industry full of volatility and regulations. Examples include the following:

  • Wealth management companies can accelerate onboarding by digitizing and automating more know-your-customer and due diligence processes, getting users to the value they want much sooner.
  • Brokerages can make their workforces more efficient by automating the new employee onboarding process so that employees can spend more time focused on customers and less time learning the mechanics of internal tools.
  • Commercial banks can automatically review loan applications, enabling quicker decisions to approve loans with less manual review.
  • Retail banks’ ability to fully automate fraud detection processes means faster response times and a reduced risk of fraud escalation. Banks can also use automation to perform eligibility checks or underwriting at the moment of data capture for instant decisions. Automated review of credit application documents can quickly help businesses to decide to approve loans with less manual review.
  • For all financial services organizations, a key use case will be leveraging AI to reduce compliance issues and thereby fines when it comes to sifting through mounds of paperwork including complaints, new account paperwork, and more.

Getting started with AI

Firms of all sizes are working to place smaller and more frequent bets on emerging technologies and channels. We’re seeing this to be especially true of AI.

As we detail in our 2018 trends report, “In Search of Real AI Answers, Brands Find Them Closer Than Expected,” experiments that begin with business outcomes in mind and work backwards to the technology are far more likely to become success stories.

With all of the possibilities that AI enables, a rapid expiration mindset combined with vigilant empathy for customers is a proven route to achieving effective AI outcomes.

Interested in knowing more?

Solstice hosted a panel of cross-vertical experts to lead an executive briefing on the current state of AI and its impact on customer experience — here are six key takeaways from the interactive discussion. You can also read about our AI practice area here.

This post was co-authored by Charlie Farmer.