"The AI Problem"
When most people think of AI in financial services, they think of automation.
If AI can read a set of financials, why not use it to open an account for a customer? Or to approve a loan application? So much time is lost in financial services to lengthy, manual processes...Surely AI is the answer?
But this is where many applications of AI in financial services stumble.
Trust is the foundation of the industry - and it compounds or collapses with every decision. A compliance misstep doesn't just result in a fine; it invites scrutiny across the board. A bad loan doesn't just hurt the P&L; it undermines confidence in the institution. These are areas where the cost of being wrong is high, and where - as a result - risk tolerances are low.
Here, existing processes are already robust and AI, no matter how high its potential, rarely gets the benefit of the doubt.
This is "the AI problem" in fintech spoken about by Simon Taylor in Fintech Brain Food: the most obvious opportunities for automation tend to be the hardest to realise. Systems are tightly regulated, deeply entrenched, and built to avoid surprises. The result? AI remains technically impressive - but commercially sidelined.
The Smarter Edge of Innovation
Does this mean that AI is stuck in a holding pattern in financial institutions? Doomed to tread water in innovation teams pending wholesale behavioural transformation?
No! There are plenty of fruitful paths for AI in financial services where the stakes are not only lower - the upside is also huge.
Think about sales enablement and account management. These are areas where AI-powered insight - not automation - creates value. Where knowing just a little more about a customer can lead to better conversations, smarter product recommendations, and higher conversion rates.
It turns out that applying AI to top of funnel activities instead of at the sharp end of product execution can yield amazing results. Generative AI is great at answering questions like:
- Which customers are most likely to benefit from a new financial product?
- Which companies in a portfolio have untapped spend that could be moved onto a virtual card?
- What recent activity signals a timely reason to check in?
...and so on. These are commercially important questions with minimal regulatory risk. If an insight is off, it's a missed opportunity - not a crisis. If it's accurate, it helps teams engage customers more meaningfully and uncover significant hidden value. The risk is low, the reward compounds.
AI Enablement, not AI Automation
The mindset shift is this: don't treat AI as an automated decision-maker. Treat it as a top of funnel collaborator.
Instead of replacing human judgment, AI can enhance it - surfacing patterns, highlighting anomalies, and suggesting next-best actions. It doesn't need to make the call on whether to write a $10 million loan. It just needs to show a relationship manager the ten customers who might benefit from that loan. This is where AI shines: not in replacing humans in high-stakes workflows, but in making them smarter and more effective. This is where AI can deliver compound benefits - without compound risks.
AI is transforming financial services - just not in the headline-grabbing ways that most people assume. It's not about autonomous credit committees or end-to-end loan processing. It's about subtly anticipating needs within a portfolio, quietly nudging sales people to act at the right time, armed with the right insight.
And in financial services, quiet innovation often turns out to be the most enduring kind.