The problem with banking’s use of artificial intelligence technology is not always in a bank’s inability or unwillingness to invest in it. In fact, more financial institutions are spending on AI and related technologies as banking leaders strive to build out customer personalization.
The problem instead can lie in a financial institution’s failure to align the technology with the bank’s strategy. As a result, a bank or credit union might buy (or build) the technology, but then abandon it shortly after or fail to follow through with it, a McKinsey report points out.
The report uses an unnamed large retail bank as an example, one that has set aside resources in the budget for machine-learning (ML) technology — one part of AI — to automate marketing and customer personalization campaigns. Sounds great, but two years later that expense still wasn’t paying off.
“The bank was still managing its personalization program much as it always had: manually and in silos,” the report states. “Although it had acquired a sophisticated analytics engine, the bank had overlooked the elements needed to turn that engine into a smoothly functioning ‘brain’. The result was a perpetual cycle of subscale efforts.”
This example is not an anomaly, the consulting firm maintains. In fact, the report says it’s commonplace in banking. However, in a time when AI investments are ramping up and becoming more critical to meet customer expectations, not taking AI or ML software seriously could impair a bank or credit union’s growth and relevance. One report from Research and Markets found the AI in banking market size globally was worth nearly $4 billion in 2020, and it predicted it will reach at least $60 billion by 2030.
A Growing Trend:
AI is becoming a salient technology in digital banking. It's forecasted to grow 15 fold from 2020 to 2030.
An American Banker survey uncovered just over half of all financial institutions (51%) report they “are actively engaging with AI and ML in pilot testing, limited use or significant use cases, with most still in the early stages.”
But, even if a bank or credit union is heavily investing in machine-learning or AI solutions and data analytics doesn’t mean they’re doing it right.
The Mistakes Banks Make With AI Investments
There is no limit to the different kinds of AI technology a bank or credit union could invest in — either to automate paperwork and back-end processes, reduce fraud or to make their customers’ lives easier on the front end.
McKinsey insists machine-learning technologies are the path forward for traditional players in banking. The growing presence of ML signals it’s an integration critical for financial institutions. Machine-learning software is generally considered a subset of artificial intelligence technology that allows computers to predict future trends and solve problems based on past data.
But, even for banks which have started investing in machine learning, there are hurdles they keep tripping over. The top five issues McKinsey highlights are:
- Sporadic and inconsistent customer data: Only 28% can rapidly integrate internal structured customer data to use in AI or ML initiatives.
- Narrow scope of machine-learning models: Only 9% have a full suite of ML models to drive personalized engagement at every customer touchpoint.
- Subscale analytics development: Only 16% of marketing teams follow a standard procedure to build and deliver AI tools, such as analytics, at scale.
- Poor campaign integration and tracking: Only 8% of businesses use insights from models in campaign execution and decision making.
- Inadequate AI risk management: Only 14% have a framework for AI governance that manages AI-related risks without compromising speed and flexibility.
Digital data aggregation alone has been a chronic pain for banks and credit unions for a number of years. But, as McKinsey points out, accumulating data in the wrong way or garnering it from the wrong source, storing it in various digital and physical locations and even over-indexing third-party data will render any AI technology useless.
Other issues play out in the integration of AI software. For instance, many financial institutions don’t take the right steps when installing the technology (which can lead to delays or ineffective results) or they fail to plug the insights gained from the technology back into a feedback loop.
Moving Forward In a Comprehensive AI Martech Rollout
Banks and credit unions cannot afford to miss out on personalizing the banking experience. Jonathan Cohen Vice President of Applied Research at AI technology company Nvidia — says that in the age of customized user experience, personalization is critical for any financial institution, specifically if they want to retain Millennial and Gen Z populations as customers.
“Once upon a time, hyper-personalization was considered nice to have. For Millennials and next-Gen Zers, slow adoption in this area could spell long-term doom,” Cohen says in an article in Global Banking & Finance Review.
There is a need for speedy, efficient and sophisticated mobile apps in addition to fortified fraudulent protections across the customer experience, he says, but all of it starts with strong AI technologies.
Five Steps to Machine Learning & AI Success
McKinsey recommends a five-tiered approach to instituting and implementing AI software, which it says has improved financial institutions’ relationships with customers.
- Identify high-value opportunities.
- Rapid activation and optimization at scale.
- Invest in fit-for-purpose martech enablement.
- Commit to creating a truly agile operating model.
- Invest in talent and capability building.
Regarding the first step, the report suggests a bank or credit union not roll out AI personalization integrations across the board with all customers. Instead, McKinsey recommends starting with a handful of “high-impact journey use cases” in a process it calls ‘opportunity identification’.
Using another unnamed retail bank as an example, the report says the bank successfully implemented AI software because its marketing teams focused on the ‘deposits churn journey’ among its affluent customers, and then moved on to the ‘investments churn journey’ for the same segment. Only when it had solved the data aggregation and customer profiles of those segments did it move to other customer segments.
“Once they were further up the maturity curve, the bank challenged its personalization teams to identify mass segment customers whose near-term value might be low but whose potential lifetime value was double or triple the average, making them good candidates to be upgraded to the affluent segment,” the report reads.
The second step cited above relates to how a bank must roll out machine-learning software throughout the company. “In addition to the advanced machine-learning ‘brain,’ rich customer-data set and marketing automation tools, organizations need a robust set of performance metrics,” the report reads.
Building a Well-Oiled AI Machine:
Integrating tech without a purpose never works. It is critical marketers apply metrics and data analysis when adding AI into the martech mix.
It will take a financial marketer time to develop these metrics, but the firm advises developing a playbook of best practices, which can shrink the learning curve.
The third step involves Marketing taking the metrics developed in the second step and running them through the machine-learning technology. This will generate campaign inputs, which can then be compared with the field data that campaign produces until a reliable customer marketing strategy takes shape.
“In addition to the advanced machine-learning ‘brain,’ rich customer-data set and marketing automation tools, organizations need a robust set of performance metrics.”
“Many companies fall into the trap of ‘throwing technology at the problem,’ but instead of an integrated tech stack, they end up with a bloated one that adds cost and complexity,” McKinsey states. “Best-in-class organizations align martech resources around their highest-priority use cases and tease apart the data, design, decisioning, distribution, and measurement dimensions they’ll need to meet their customer goals.”
The fourth and fifth steps deal with team growth. The marketing team cannot be the hand carrying the whole weight of machine-learning and AI software integrations.
“In the bank’s case, the marketing team was tasked with managing the program, but marketers had only sporadic access to analytics resources, forcing them to fall back on basic heuristics that were easier to manage but less effective at personalization,” McKinsey states. “As a result, the institution struggled to meet its retention and satisfaction targets for key segments.”
The Role of Bank Marketers in AI Integration
While the back-end side of machine-learning software will naturally be highly technical and data-heavy, it is important marketers remember the front end should always be simple and easy to digest.
Ultimately, it is the job of the financial marketer to be the translator between the two.
Bank marketing teams can adjust phrasing used behind-the-scenes to clarify information for the customer. For example: Instead of using sales messaging such as “Try our zero percent introductory rate,” a bank or credit union could use what McKinsey calls journey-based communications, such as “Here’s how to make your holiday stress free.”
McKinsey also recommends adjusting the tempo of marketing communications based on findings from the ML technology. “Instead of monthly or quarterly campaign releases, they shifted to a daily or weekly tempo. Other companies can do the same.”