logo

Klarna’s AI revolution: The future of an AI-driven bank

https://live.theasianbanker.com/

The financial industry is undergoing a rapid transformation, with banks and fintechs integrating artificial intelligence (AI) to improve efficiency and customer experience. While some firms merely add AI-driven features, Klarna is taking a far more radical approach—restructuring its entire organisation around AI to become the most efficient AI-driven bank.

Klarna, long known as a leader in buy now, pay later (BNPL), is a fully licensed Swedish bank. In 2023, CEO Sebastian Siemiatkowski announced that Klarna aimed to become “the most efficient AI-driven bank.” A year later, Siemiatkowski announced on X the significant strides toward this vision:

  • Workforce reduction by 50% – Traditional roles have been replaced with AI-driven processes.
  • Elimination of 1,200 SaaS tools – Major enterprise platforms, including Salesforce and Workday, have been phased out, consolidating operations into a unified AI-powered ecosystem.
  • Proprietary AI infrastructure – Off-the-shelf software has been replaced with in-house AI systems optimised for automation, data standardisation, and efficiency.
  • AI literacy across the workforce – Employees are trained to incorporate AI in decision-making, embedding AI into Klarna’s corporate DNA.

While these changes enhance efficiency, they also raise operational risks, such as reliance on internal AI models that must continuously evolve to keep pace with industry standards.

AI at Klarna’s core

Unlike traditional banks that integrate AI into existing structures, Klarna is rebuilding its operations from the ground up. Key initiatives include:

  • Unified data architecture – AI standardises and centralises knowledge across teams, eliminating silos.
  • AI-driven decision-making – Machine learning models optimise fraud detection, risk assessment, and customer service.
  • Graph databases for business intelligence – Technologies such as Neo4j consolidate disparate data sources, enhancing operational insights.
  • Automation in customer service – AI-powered chatbots and decision engines cut costs while improving speed and accuracy.
  • AI-driven risk management – Klarna’s AI models predict default risks, analyse transaction patterns, and optimise lending decisions in real time.

However, scaling AI-driven decision-making comes with regulatory challenges, as financial authorities worldwide—particularly in the European Union with the Artificial Intelligence Act and in the United States with the Colorado AI Act—are increasing scrutiny on AI’s role in risk assessment and consumer protection.

Eliminating SaaS to enhance efficiency

Klarna’s decision to phase out enterprise  software as a service (SaaS) platforms reflects a growing trend in financial services—moving away from fragmented vendor ecosystems toward streamlined, AI-driven solutions. Instead of relying on multiple third-party tools, Klarna is consolidating its infrastructure to increase efficiency, reduce duplication and enhance security.

The challenge with excessive SaaS reliance is not just cost. It introduces:

  • Fragmentation – Too many vendors create disjointed workflows and information silos across the organisation.
  • Duplication and redundancy – Overlapping functionalities across different tools increase inefficiencies rather than solving them.
  • Security risks – Data passing through multiple third-party systems increases exposure to security vulnerabilities.
  • Interoperability issues – SaaS providers follow different standards, protocols and  application programming interfaces (APIs), leading to integration difficulties.
  • User complexity and training costs – Teams must navigate multiple interfaces and undergo vendor-specific training, slowing adoption.

These concerns align with broader industry shifts. According to NVIDIA’s “State of AI in Financial Services: 2025 Trends”, fewer institutions are identifying additional AI use cases (a 10% decline), but engagement with third-party AI partners has increased to 32%, suggesting firms are seeking external AI expertise. On the talent front, AI hiring has increased by 8%, but remains a small growth, indicating that in-house AI development is still in its early stages.

Further supporting this trend, the Digital Banking Report (August 2024) by Jim Marous for OpenText reveals that 69% of banking organisations prefer to buy and deploy third-party AI solutions, while only 16% build internally, and 15% partner with technology providers. This suggests that, despite the push for AI, most financial institutions still rely on external vendors.

By moving to proprietary AI infrastructure, Klarna aims to:

  • Improve data privacy and security – On-premise AI models provide greater control over sensitive information.
  • Enhance operational agility – AI-driven workflows allow faster adaptation to regulatory changes and customer needs.
  • Reduce long-term costs – While developing in-house AI is expensive upfront, it eliminates ongoing SaaS licencing fees.

As AI adoption grows, the financial sector will likely continue shifting from SaaS-heavy tech stacks to AI-driven ecosystems, particularly as firms seek more unified, scalable and secure digital infrastructures. Klarna’s approach may be ahead of the curve, but the long-term impact remains to be seen.

The Pareto Principle: Building only what matters

Klarna’s AI strategy for its internal tools follows the Pareto Principle—80% of users rely on just 20% of software functionality. Instead of paying for bloated SaaS solutions, Klarna is developing:

  • Lean, purpose-built AI systems
  • Only the core functionalities needed for efficiency
  • Scalable tools that evolve with Klarna’s business
  • AI-powered automation that reduces IT maintenance costs

While this approach enhances agility, it also creates potential downsides, such as the burden of maintaining proprietary software. Unlike SaaS providers that continuously update their systems, Klarna must sustain, refine and secure its AI ecosystem internally.

The future of AI-driven banking: Buy vs Build

Klarna’s transformation highlights a key debate in banking: should firms buy third-party tools or build proprietary infrastructure? While enterprise SaaS remains useful for many financial institutions, Klarna’s shift suggests that AI makes in-house development more viable than ever.

For banks evaluating AI strategies, key questions include:

  • Efficiency – Can AI eliminate redundant processes and manual workflows?
  • Compliance – Can proprietary AI systems meet regulatory standards?
  • Scalability – Do AI-driven solutions support long-term growth without excessive maintenance costs?
  • Competitive advantage – Will in-house AI provide an edge over firms relying on external vendors?

Klarna’s AI-first future

Klarna’s transformation is not just about cost-cutting—it redefines what it means to be an efficient digital bank. By eliminating legacy software and embedding AI into every aspect of its operations, Klarna is setting a potentially industry-shifting precedent.

However, whether Klarna’s AI-first model will serve as a blueprint for the industry—or an outlier—remains to be seen. As AI reshapes financial services, the question for banks is no longer whether to adopt AI, but how to integrate it responsibly and effectively.

Dom Monhardt is the founder of one-fs.com, a leading fintech and digital banking newsletter in the Middle East and North Africa (MENA).