AI Investment in Banking: Why Cost Efficiency Isn't Enough and How to Build Self-Funding AI Programs.
Banks are rushing into AI; but who’s paying for it?
Artificial Intelligence, (AI), has moved from buzzword to boardroom priority at a pace rarely seen in previous technology waves, something that is very exciting to see! Across global banking we are seeing investment in AI is accelerating. Globally, banks are approving big budgets, teams expanded, vendors onboarded, and pilots launched, often simultaneously.
And to be clear, this momentum is not misguided. The expansion of AI brings genuine and transformative benefits. From faster decision-making and improved customer experiences to enhanced risk management and productivity gains, AI has the potential to fundamentally improve how banks operate and compete.
But there is a growing issue hiding in plain sight: many banks are investing heavily in AI without a clear, sustainable plan that determines how those investments will pay for themselves.
From operational wins to strategic risk
In recent times, banks have deployed AI in relatively contained, operational areas. Think repetitive tasks such as fraud management, credit scoring, customer service chatbots, document processing, and process automation. These use cases were attractive because they were measurable, lower risk, and focused on efficiency. The return on investment was typically framed around cost reduction, headcount optimisation, or error reduction. That phase is now largely complete.
Today, banks are pushing AI deeper into the strategic and fundamental core of their businesses. AI is being used to influence pricing, customer engagement strategies, product design, balance sheet optimisation, marketing effectiveness, and even capital allocation decisions. These are no longer back-office enhancements; they are front-office, revenue-impacting, and brand-defining capabilities. This is hugely significant because it dramatically alters the risk profile.
As AI moves closer to the heart of the bank, the challenges multiply. First, there is strategic risk. When AI informs decisions around pricing, cross-sell, customer prioritisation, or credit appetite, small model biases or data quality issues can have outsized commercial and reputational consequences. A flawed recommendation engine does not just reduce efficiency; it can erode trust, damage customer relationships, or distort long-term value.
Second, there is financial risk. AI investment is expensive and ongoing. Infrastructure, data engineering, model development, governance, talent, compliance, and vendor costs do not taper off after launch, they compound. Yet many AI business cases are still framed around vague future benefits or “keeping up with the market,” rather than hard, incremental revenue or profit outcomes.
Third, there is execution risk. Strategic AI requires deep integration across data, systems, and teams that were never designed to work together. Many banks are discovering that deploying AI at scale is less about algorithms and more about organisational alignment, decision ownership, and cultural change areas where progress is often slow.
Finally, there is opportunity cost. Funds invested in AI is capital not deployed elsewhere. When investment decisions are made reactively or “blindly,” without a clear path to value creation, banks risk building impressive capabilities that do not materially move the bottom line.
The question not enough people are asking: Where is the return on this AI investment?
What is striking is how rarely AI investment is explicitly linked to incremental revenue growth.
Cost efficiency remains the dominant narrative, but efficiency alone cannot justify the scale of investment now underway. At some point, boards and shareholders will ask a simple question: how is AI helping us grow?
This is where many banks are currently exposed. They are funding AI programmes out of existing budgets, transformation funds, or discretionary spend, without a clear mechanism for those investments to generate new income streams or unlock latent value already sitting within their customer base.
Consider a smarter way to fund your AI investments
At Profit Insight, we believe AI investment should not be a cost burden that banks simply absorb. It should be at least partially self-funding.
Banks already hold vast amounts of customer, transactional, and behavioural data that can be used, responsibly and compliantly, to uncover incremental revenue opportunities. These opportunities often sit between products, channels, and customer segments, unnoticed by traditional analytics but visible when data is connected and insight is activated.
The most effective AI strategies are those tied directly to revenue impact: improved customer engagement, smarter cross-sell, reduced attrition, better pricing decisions, and more personalised experiences that customers actually value.
Our role is to help banks identify and realise these incremental profits, creating tangible financial returns that can directly contribute to funding wider AI investment. Turning the question of “How much will AI cost us?” to “How much value can AI unlock for us?”
These thought came from some really interesting conversations with clients and partners in this space recently so if it is something you are experiencing, or you are interested in finding out more, please visit our website www.profitisight.com or contact me or any of the Profit Insight team and let’s explore how we can help your bank fund these projects.
