March 5, 2025 | By Ben Tannenbaum

How AI is Reshaping Enterprise Software: Lessons from Klarna's SaaS Revolution

When Klarna CEO Sebastian Siemiatkowski mentioned during an investor call that the company had shut down Salesforce and approximately 1,200 other SaaS tools, it sparked a firestorm of speculation. Was this the beginning of the end for enterprise SaaS? Had AI made traditional software obsolete? The reality, as he recently clarified in a detailed Twitter thread, is both more nuanced and more profound.

Klarna's journey began with a fundamental realization: enterprise knowledge was hopelessly fragmented across dozens of SaaS platforms, each with its own data model, interface, and approach. Documents scattered across document management systems, projects divided between ticketing and kanban boards, analytics split between various dashboards, customer information siloed in CRM systems, and employee data isolated in HR platforms. This fragmentation created what Siemiatkowski calls "an unnavigable web of knowledge that required a tremendous amount of Klarna-specific expertise to operate and utilize."

Early AI experiments at Klarna revealed a crucial insight. While many were excited about the ability to "feed all your PDFs, all your data sources to an LLM," the Klarna team quickly discovered that the old data science principle still applied: "garbage in, garbage out." As our own research at Aiso has shown, AI systems are only as good as the data they're trained on. Our analysis of millions of ChatGPT conversations reveals that users are increasingly seeking high-quality, contextual information for important decisions — with 15.2% of transactional queries focused on automotive purchases and 14.8% on computer hardware.

Rather than simply feeding fragmented data into AI systems, Klarna took inspiration from Wikipedia's collaborative knowledge graph and partnered with Neo4j to build a unified knowledge system. This approach focused on identifying truly valuable data, eliminating duplication and contradictions, creating connections between previously siloed information, and standardizing access, versioning, and audit capabilities. The result wasn't replacing SaaS with AI, but rather creating a unified knowledge foundation that AI could effectively leverage.

The real breakthrough came when Klarna connected this unified knowledge graph to their internal AI systems. Using tools like Cursor AI, they could rapidly deploy new interfaces and interactions with their knowledge base. This mirrors what we're seeing in consumer AI usage. Our research shows that users engage in multi-turn conversations about complex topics, with regional variations in responses. For example, automotive queries in the US surface information about Tesla and Ford, while European users see content about Volkswagen and Rimac. For a deeper dive into these regional variations and industry-specific search patterns, check out our detailed analysis What Are People Searching for on ChatGPT?

Siemiatkowski's prediction for the future of SaaS is particularly insightful:

1. Consolidation, not elimination: Fewer SaaS providers will dominate the market
2. Knowledge hubs, not point solutions: Successful platforms will integrate knowledge across domains
3. Opinionated software wins: SaaS that maintains strong opinions about best practices will outperform generic databases

This aligns with our observations of how AI is reshaping consumer behavior. Users aren't just seeking information — they're seeking contextual guidance through complex decisions.

The key takeaway from Klarna's experience isn't that AI will replace enterprise software, but rather that companies must become "AI-first" in their approach to knowledge management. This means:

1. Breaking down data silos
2. Creating unified knowledge graphs
3. Focusing on data quality over quantity
4. Building AI interfaces that leverage this knowledge foundation

What Klarna's experience demonstrates is that AI isn't simply a new feature to bolt onto existing systems — it requires a fundamental rethinking of how organizations structure and access their knowledge. As our research into ChatGPT usage patterns shows, AI excels at providing contextual, conversational guidance through complex decisions. But this capability depends entirely on having high-quality, well-structured information to draw from.

The companies that thrive in this new era won't be those that simply adopt AI tools, but those that rebuild their knowledge foundations to make AI truly effective.

For more on how AI is reshaping consumer behavior and what it means for brands, visit our detailed analysis of ChatGPT search patterns.