Maximizing AI Payoff in Banking Will Demand Enterprise-Level Rewiring
Sprinkling AI around a bank like magic fairy dust, a little here, a little there, will produce only little results, according to a new report by McKinsey. Instead, major commitments to rebuild banking organizations from the ground up with artificial intelligence will produce meaningful improvements.
By Steve Cocheo, Senior Executive Editor at The Financial Brand
When it comes to technological change, banks can be notorious toe-dippers, trying out innovation here and there. But a new McKinsey report advises that this cautious approach when it comes to artificial intelligence won’t produce the cost savings and process improvements that the industry needs to regain productivity and compete more effectively.
Instead, in the consulting firm’s view, banks have to be bold, setting out to reimagine entire sections of the organization as they could function using multiple types of AI. This mindset is not unlike poker players going "all in" on a hand that they really believe in.
As reported recently on this website, there’s been a backlash to the hype around GenAI and other AI tools in some quarters in banking. McKinsey’s own global AI survey has found that while adoption of some AI in organizations has grown globally, "breadth of adoption — measured by the deployment of AI across multiple enterprise functions — remains low, and many organizations are still in the experimental phase."
You can’t convert an entire institution into an AI-first organization all at once. But McKinsey advocates tackling the challenge of AI adoption and transformation domain by domain, addressing change in the natural large clusters that banks operate through.
"Some banks have taken a ‘Let 1,000 flowers bloom’ approach to AI adoption," says McKinsey’s Carlo Giovine. "They have given access to AI tools to many of their employees and then asked them to solve small problems with those tools. They then apply AI to solving the small things they do on a daily basis."
The result is that banks may gain small time savings but these are isolated and don’t connect with each other, according to Giovine, a London-based partner who works at QuantumBlack, McKinsey’s advanced analytics and machine learning center.
Even a mega institution with lots of chips to spread around sees the need for concentrated efforts. In fact, in a recent interview, JPMorgan Chase’s Teresa Heitsenrether, EVP and chief data and analytics officer, said innovation needs directing. Her comment in that interview echoed Giovine’s: "You want to avoid the proverbial ‘thousand flowers blooming.’ You want to make sure that you’re focusing on the things that are going to add real enterprise value."
Producing Real Returns for AI Investment Means ‘Domain-Level’ Thinking
So, what’s the alternative to many banks’ piecemeal, siloed approach to AI?
"You have to take an end-to-end point of view at each domain," Giovine says in an interview with The Financial Brand. This means more than simply being willing to adopt AI tools. It requires "having a hard look at your process and rewiring it with AI technology."
McKinsey estimates that most banks consist of around 25 domains. Among those where AI transformation could be most effective, the firm says, are underwriting, customer acquisition, frontline sales enablement, relationship management, and management of self-service banking channels.
Policies may need adjustment, process flows must be reconsidered, and even the frameworks used to evaluate business risks may need rethinking. As McKinsey’s report illustrates, the task encompasses business thinking as well as technological expertise. One can’t proceed without the other. In fact, in budgeting for such change, Giovine says $4 out of every $5 will likely go for change management, and $1 for the actual technology.
"When you take a domain view, you avoid falling into the trap of saving a little here and a little there," says Giovine. Effective AI adoption requires more than me-too efforts at taking up GenAI for this project or that, just for bragging rights.
Beyond thinking in broad strokes of AI’s applicability in the bank, McKinsey holds that an institution has to be ready to adopt multiple kinds of AI set up in a way to work with each other. This includes analytical AI — the types of AI that some banks have been using for years for credit and portfolio analysis, for instance — and generative AI, in the forms of ChatGPT and others, as well as "agentic AI."
In general, agentic AI uses AI that applies other types of AI to perform analyses and solve problems as a "virtual coworker." It’s a developing facet of AI and, as described in the report, is meant to manage multiple AI inputs, rather than having a bank lean on one model.
Overall, according to the report, "These agents, like humans, have the capacity to eventually be able to plan (for instance, organize a workflow encompassing a series of tasks), think (come up with chain-of-thought reasoning), and act (use digital tools)."
Read more: 8 Mistakes That Will Guarantee AI Fails at Your Bank
Making Domain Thinking Digestible for AI Adoption
Giovine says once the bank has figured out what AI strategy will animate the domain it is tackling, it can break the challenge into smaller chunks — part of an overall effort, not piecemeal.
"In my experience, you take a multi-step approach," Giovine explains. "It starts with taking one meaningful slice of the domain. It could be one business segment. Or one set of cases."
Then the work begins, with a quarter or so to build a pilot and begin testing it. Metrics are critical.
"You measure the outcomes you want to achieve and at the end of the pilot you will typically come out with a very good understanding of how to scale it," Giovine says. Over six to 12 months after the pilot, "you can scale it over a good chunk of the domain."
And here, the consultant says, is where the bonus kicks in: Often a good deal of the work done to bring AI thinking to one domain can be re-used. This applies to both the business thinking and technology.
"You could train a new set of users from a different domain," says Giovine. "The learnings and the things that you build become a real asset for the bank."
Read more: Why the Power of GenAI Lies in the Augmentation, Not Automation (or Replacement), of Bankers
Is Domain-Level Thinking Useful for Banks Beyond the Giants?
A natural question is, how far down the banking food chain will this approach work?
"I would say it’s equally valuable for all sizes of bank," says Giovine. He says big institutions have more complex data and greater information technology and software complexity. This requires greater customization of the AI tools to make it work. More standardized approaches can still take smaller players a good way down the road, according to Giovine.
ICYMI:
