Artificial Intelligence: How to Move from Science Project to Value Add for Banks

“What we’re seeing around AI… it’s still a large and very expensive, maybe even the largest and most expensive science project that we’ve ever endeavored on in humanity. There’s still a lot to be proven,” said James Dotter, MX’s Chief Business Officer. 

There is a lot of opportunity and investments — and hope — being put into artificial intelligence. IDC estimates worldwide spending on technology to support AI strategies will reach $337 billion in 2025 — and more than double to $749 billion by 2028.

But can AI live up to these expectations and investments? Or is it just an expensive science project? McKinsey’s latest global survey on AI shows that while adoption has increased significantly across organizations and industries, the breadth of adoption remains low and many organizations are still in the experimental phase. 

While AI can help create personalized customer experiences,streamline inefficient processes, and uncover new growth opportunities, it’s only as good as the data that fuels it. To succeed with AI, financial providers need to prioritize getting their data strategy right first.

Starting off with the Right Data

Too often, AI users find themselves hung up on the outcomes of AI and neglect the starting point. At Money Experience Summit 2024, Dr. Leda Glyptis said, “I care about the organizational hang ups that will always forever make the job an uphill struggle no matter how intelligent you get about AI.”

Using AI for the sake of AI will not solve operational efficiency problems until teams can assess their starting point. That starting point is data.

Here are two areas where financial institutions and fintechs can focus first to ensure quality data to fuel their AI use cases: 

1. Break Down Data Siloes

The average consumer has at least 5 or more financial accounts with various banks, credit unions, and other financial providers. This disjointed money experience leaves consumers — and the financial providers who serve them — with only part of the picture. 

Without connecting those various accounts and sources of financial data together, organizations face significant blindspots when it comes to delivering the right products and services to meet consumer needs, targeting consumers with the right messages, and accurately predicting and preventing churn. 

This means AI systems will face the same blindspots, leading to inaccurate data and even the potential for AI hallucinations. In fact, popular AI models like ChatGPT and Google’s PaLM make up facts anywhere from 3% to 27% of the time, depending on the tool. And, 89% of machine learning (ML) engineers who work with generative AI say their models show signs of hallucination. 

2. Transform Raw Data into Clear Insights

Poor data quality can be costly for businesses. Vanson Bourne and Fivetran found that models trained on inaccurate, incomplete, and low-quality data caused misinformed business decisions that impacted an organization's global annual revenue by 6%, or $406 million on average.

In the financial industry, the stakes are even higher — unclear, raw data doesn’t just impact businesses. It can negatively affect customers. For AI, it means garbage in, garbage out. 

Raw transaction data is messy and confusing — so outputs from AI tools based on this data will likely also be messy and confusing. No one can take action based on random strings of characters. Enter data enhancement, which transforms raw, chaotic transaction data into enriched, structured, and actionable insights.

MX’s Chief Advocacy Officer Jane Barratt said, “You can’t jump over a lack of data strategy and go straight to AI.” Businesses that want to excel in their AI adoption and stand out from the competition must rethink their data strategy — from data collection through implementation. 

Want to read more about our take on AI in 2025 and other top trends? Access our 2025 predictions report.