Why 80% of Banks Lag in AI Deployment: GenAI Finance Adoption 2025
- Saveliy M.
- Sep 27
- 5 min read
Updated: Oct 26
Summary
A recent industry survey found only a small fraction of banks have fully deployed generative AI.
The promise is high – banks see ~20–40% efficiency gains in early AI projects.
Leading banks highlight isolated successes.
Major banks (JPMorgan, HSBC, etc.) have launched GenAI pilots and internal tools, aiming to improve customer service and back-office work. At the same time, smaller banks cite resource and talent shortfalls[5]. Overall industry spending on AI (in banking and fintech) has surged: financial firms reportedly increased AI budgets by double-digit CAGR in recent years[8], but converting pilots into live deployments has been slow.
Legacy technology is a core issue. Industry analyses note that “47% of banking workloads run on COBOL” (an older programming language)[3]. Such monolithic systems make real-time AI integration difficult, forcing workarounds or long upgrade timelines.
Regulators and auditors also influence pace. New frameworks like the EU’s AI Act and anticipated US guidelines are pushing banks to build explainability and fairness checks into AI systems, adding compliance overhead[6]. Many banks report that regulatory and data-security requirements rank among their top AI concerns[6].
Details of the Development
The headline claim that “80% of firms have no formal GenAI in finance” reflects industry analysts’ observations of slow adoption. In practice, dozens of banks have only proof-of-concept AI projects. Historical patterns suggest the upside is large: McKinsey has estimated that AI could cut banks’ operational costs by ~30–40% if fully adopted, especially in middle- and back-office functions.
Pilot results vary. In one case, a large bank’s AI coding assistant increased developer output by about 40%[2]. Other pilots in fraud detection and compliance are said to reduce errors and workload by ~20–30%. However, many trials yield minimal ROI. Reports indicate billions spent on AI initiatives so far with mixed results, partly because legacy data pipelines and incomplete staff training limited their scope.
Talents and tools: Surveys find roughly 70% of banks report gaps in AI/data science talent[5]. Banks are hiring or retraining staff, but the industry shortage means many smaller players lag. Meanwhile, fintech and tech firms are bundling AI into SaaS platforms, enabling “plug-and-play” use cases that avoid legacy entanglements. For example, some digital banks already credit AI-driven personalization with higher customer engagement.
Market impact: If current trends persist, experts warn that banks delaying AI could lose market share to fintechs. Digital-first firms have been able to adopt new tools faster; one estimate is that “digital-native” finance companies have captured ~15% more market share in relevant segments than incumbents over the past two years (though we did not find a public source for this exact figure). In any case, the gap between early AI adopters and laggards appears to be widening.
Industry Reactions to Adoption Challenges
Bank executives acknowledge pressure. DBS Bank (Singapore’s largest lender) publicly noted that its focused AI projects generated hundreds of millions in benefits. In 2023 DBS reported about S$370M in AI-driven savings and revenue enhancement[7]. This is often cited as proof of concept for cautious adopters.
Analyst advice is to start small. Consulting firms and fintech analysts suggest banks begin with low-risk pilots (e.g. compliance document review, anti-money laundering tools) before moving to customer-facing GenAI. Many regulators (Fed, ECB) have also stressed that they encourage innovation with safeguards – they are watching how banks build AI governance, rather than broadly rejecting the technology.
Competition and partnerships: Non-bank competitors see opportunity. Firms like Revolut and Stripe have integrated AI features (e.g. chatbots, automated fraud alerts) and tout higher customer retention, implying a roughly 10–20% edge from AI services. Tech providers (IBM, Google, AWS) are offering hybrid-cloud platforms to help banks run AI alongside legacy systems. None of the major banks are reportedly resisting AI outright, but most emphasize the need for phased rollouts.
Market outlook: Leading consultancies still project a large finance AI market. For example, a 2024 Deloitte report (no single number found) suggests global financial services AI could approach the low hundreds of billions by 2029. The consensus view is that early adopters will gain share and valuation, while strict laggards risk falling behind.
Financial Analysis of Banking AI Using Key Methods
To assess the impact of GenAI on finance, we outline some common approaches:
Trend Extrapolation (Growth Projection)
Looking at past AI investment growth, financial industry spending on AI tools has been rising at double-digit rates. A hypothetical “base case” might assume ~15–20% annual growth in finance AI adoption, leading to, say, ~$200–300B market by 2027.
In a bull scenario (regulatory easing, high demand) we could see 25–30% CAGR, while a bear scenario (security issues, slow pilot success) might be only 10%. In practice, even the base case would imply that banks adopting AI early could boost revenues and cut costs by a noticeable margin (5–15%), while lagging banks might lose a few percent of share.
Discounted Cash Flow (DCF) Estimates
We illustrate using two representative banks. For an early adopter (e.g. JPMorgan) assume new AI services add $2 B in revenue by 2026 at 20% margin (so $400M FCF incremental per year). Discounted at ~10% WACC, over five years this yields roughly $2.5B in present value, with a terminal value (assuming 12× FCF) of ~$30B.
If JPM’s shares were 350M outstanding, that’s ~$50–60/share uplift.For a slower adopter (e.g. Bank of America), assume only $500M incremental revenue with 20% margin ($100M FCF/year).
Discounted similarly, that might be only ~$0.8B PV and ~$10B TV, implying ~$10–15 of stock lift. These back-of-envelope DCFs suggest early adopters could see outsized EPS growth; in other words, AI success could warrant higher P/E multiples (see below).
Comparable Multiples
Banks are often valued at ~10–12× trailing EPS. If AI drives, say, a 5–10% EPS increase for adopters, their forward P/E could reasonably expand by a few points. For example, JPMorgan now trades ~11×, but if the market believes AI will sustainably boost earnings, analysts might justify raising it toward 14×. Comparables in fintech (PayPal, Adyen) trade much higher (~20–30×) reflecting growth. Thus we estimate a bull case of ~12–14× for AI-savvy banks, vs. ~8–10× for laggards.
Scenario Analysis
Optimistic: 50% of banks deeply adopt GenAI by 2027 – global finance AI market ~$250–300B, driving 20–30% faster productivity for adopters. Early adopters gain 10–15% market share.
Base Case: 30–40% adoption by 2027 – AI market ~$200B, ~10–20% productivity lift. Leading banks hold or slightly gain share, laggards shrink modestly.
Pessimistic: 20% adoption by 2027 (many projects stall) – market ~$150B, only 5–10% efficiency gains. Laggards risk losing ~5% market share to fintechs.
Investor Implications and Predictions
Our rough recommendations (assuming no contraindications): Buy JPMorgan (JPM) and Citi (C) for their early AI initiatives; these could earn ~+30–50% ROI over 12–18 months if GenAI drives profits. Longer term, Fintech/AI ETFs may capture broad gains, but carry higher risk.
We forecast the finance AI market could reach ~$200B by 2027 (roughly +20% CAGR), and that cloud/tech stocks powering AI (e.g. Nvidia, MSFT) will also benefit. Key risks include high implementation costs, talent shortages, and regulatory pushback – monitor quarterly earnings guidance and central bank tech task forces.
Investment Vehicle | 12-Mo Return Est. | Risk Level | Rationale |
JPMorgan (JPM) stock | +30% | Low | Leadership in banking AI |
Citi (C) stock | +20% | Medium | AI-enhanced fraud detection gains |
Fintech/AI ETFs (e.g. ARKF) | +25% | High | Broad AI adoption in fintech |
Disclaimer
This analysis is for informational purposes only and is not financial advice. The numbers and scenarios are based on publicly available data and estimations; outcomes may differ materially based on regulatory decisions, market behavior, or other developments, which could extend beyond 2025. Always consult a qualified financial advisor before making investment decisions.
Join Our Community of 60,000 Innovators
Stay updated on the latest news and insights!
Instagram: Instagram.com/aspirescapital
Subscribe for More Updates: aspirescapital.com/signup
Comments