Part 3: Operating Efficiencies is the 3rd installment of Flagship’s multi-part series assessing the potential impacts of AI on payments and fintech. Read Part 1 & Part 2
If you are an investor in fintech, AI gets exciting when you consider the potential for efficiency gains and resulting earnings acceleration. In a decade, financial services will be almost entirely digitized, powered by AI. Most of us reading this article already consume our financial services digitally. But in the background, financial services infrastructure and operations remain frustratingly manual and people dependent, including costly deployment of people for rote tasks. Fintechs, leveraging AI, will innovate all aspects of financial services operations, transforming the industry and unlocking shareholder value while allowing people to focus on higher-value purposes.
The Ethics of AI and Jobs Displacement
Before we discuss the financial upside of displacing people with machines, let us first address the ethical sensitivities of the topic. For centuries, skeptics have preached the downsides of innovation, “machines replacing people is bad.” History proves otherwise. Innovation created more jobs than it displaced across human history. Innovation is at the heart of society’s ability to support larger populations at higher standards of living. As shown in Figure 1, U.S. jobs grew at 2% per annum from 1950 to 2025, a period of tremendous innovation. Employment never declined for more than two years in a row during this period (declines driven by macroeconomic cycles and more recently by COVID). Labor force participation increased from 55% to 60% from 1950 to 2025. Although we note the deceleration of jobs growth since the rise of the internet and web-based technologies (from 2.2% to 0.8% jobs growth), we generally dismiss the idea that innovation is fundamentally bad for people.
AI will not be good for all people either. AI will create labor market disruption, forcing people to find new forms of employment. Will AI lead to job reduction in financial services in the aggregate? … most likely yes. Product innovation (e.g. mobile banking apps) tends to create net new jobs, while process innovation (e.g. chatbots) tends to remove jobs. We expect AI to deliver more process innovation than product innovation in financial services, although both will happen. Most financial services investors would agree that eliminating jobs and people costs is a necessary evil in progressing towards a more sustainable industry.On the precipice of AI disruption, we expect financial services jobs to decline in the aggregate in the coming decade. You need only to read the headlines of major global banks to know that jobs reduction is likely. However, we expect jobs growth in fintech as a sub-segment, although fintech has its own people cost challenges. AI disruption will favor leading fintechs, who will leverage both people and AI development to power ongoing financial growth. Jobs displacement will be more severe in financial services companies that do not adapt to AI.
Even Fintech Has Lots of Room for Efficiency Improvements
Financial Technology, by definition, should be a high-leverage industry where topline growth consistently and materially exceeds expense growth. Fintech providers should excel at building technology that dramatically improves operating leverage. When we examine the reality, however, we see an industry achieving middling results relative to the potential. Figure 2 charts a sample of fintech companies by one simple metric: Gross Profit Growth vs. Operating Expense Growth (which we refer to as our efficiency growth index). A company delivering positive efficiency gains has an index over 1.0. Our entire sample of fintech companies delivered a weighted average efficiency growth index of 1.5 from 2021 to 2023. Illustratively, this 1.5 index equates to 15% gross profit growth with 10% opex growth, fueling a 20% operating income growth. These kinds of growth and efficiency gains are decent, but far from the potential upsides if you are bullish on AI.
AI will help fintech leaders to improve and accelerate efficiency gains. This is not to say that all fintechs should be operating well above 1.0 on this efficiency index. It is understandable that fintechs investing in growth acceleration may operate below an efficiency growth index of 1.0 for some time. But all fintechs should have efficiency gain objectives on the horizon. To be a winner in fintech, we believe that you must use AI to increase efficiency performance in the coming years. We expect the fintech industry overall to materially improve its efficiency growth index from 1.5 in the coming years due to AI. Klarna, which we discuss as a case study shortly, is effectively off-the chart on this metric based on their all-in investment in AI.
People Cost Are the Key Operating Constraint in Fintech
People costs is the reason why the fintech industry is not yet delivering greater efficiency gains. As shown in Figure 3, personnel costs make up the majority of operating expenses in the industry, representing approximately 60% or more of the total operating costs for the fintechs examined in the sample.People are expensive. As Figure 4 illustrates, the average cost per FTE for fintechs ranges from $60k per FTE per annum to more than $200k (the range mostly driven by the global location of the people). People are also only getting more expensive, particularly the most talented. Figure 4 admittedly includes a small, random sample of fintechs admittedly, but most of these fintechs experienced avg. FTE cost inflation of 8% or more from 2022 to 2023.
The fintech industry has a tremendous opportunity to leverage AI to get more out of people, to grow revenue and profit while slowing or even reversing people cost growth. This is not an outright call to start whacking people from your organization. Rather, it’s about being smart investing in both people and AI.
AI Will Drive Operational Improvement Across Many Functional Domains
So, how will AI allow fintechs to accelerate efficiency gains? As we describe in Figure 5, the AI efficiency opportunities are broad, though we expect the biggest financial impact in the technology and servicing costs. Fintech companies invest massively into ongoing development and maintenance of software and related technologies. AI will help to power software development and testing, allowing people investment to focus on feature and quality improvement. Similarly, despite the moniker, many fintech companies still power front and back-office operations with people and manual processes. AI will eliminate most (not all) of these manual functions in the coming decade.
Klarna Blazing The Path on AI
Klarna is a unique and powerful case study on the ability of AI to power efficiencies within fintech. Klarna is on the bleeding edge of AI possibilities, leveraging AI-based tools such as OpenAI to change all aspects of its operating model while dramatically reducing costs and people dependencies along the way (as shown in Figure 6).
Klarna uses AI across the organization, including:
- Underwriting – Klarna is a seasoned user of AI to underwrite consumers based on behavior, helping to make smaller and more transactional credit available to consumers.
- Marketing – Klarna uses AI to develop marketing content such as image generation. AI helped to reduce Klarna’s spend on external marketing agencies by 25%.
- Service Assistance and Automation – Klarna’s AI Assist, a chatbot written in collaboration with OpenAI, handles two-thirds of all customer service inquiries. According to Klarna, AI was already doing the job of 700 service agents in Feb.’24.
- Accounting and HR – Klarna announced plans to jettison software such as Salesforce and Workday in favor of customer, AI-powered solutions, saving millions as a result
Conclusions
Not every fintech is as bold as Klarna, a company all-in on AI in the lead-up to its pending IPO. But Klarna does illustrate the art of the possible in the fintech industry, an industry with massive opportunities for acceleration of operating efficiency. For fintech winners, AI will change the game, allowing ongoing growth while reducing people constraints. This is not to say that people don’t matter in fintech. People will continue to lead fintech to new heights, but these people will also be better directed to higher-value pursuits, while AI does the rote work of answering service inquiries and writing software code.
In our next piece in this AI series, we will focus on AI’s impact on UX and product improvement and revenue-driving ability; admittedly, this is a more optimistic topic than cost and job reduction.
Please do not hesitate to contact Joel Van Arsdale at Joel@FlagshipAP.com or Alessandro Mighetto at Alessandro@FlaghsipAP.com with comments or questions.