AI Commoditizes Code and Data Switching Costs, Reshaping Software Moats

The Looming Disruption: How AI and Data Will Reshape Software and Create New Moats

The conversation between Sebastian Siemiatkowski, CEO of Klarna, and Harry Stebbings on The Twenty Minute VC podcast reveals a profound, yet often overlooked, shift in the software landscape. While the cost of generating software plummets due to AI, the true disruption lies in the diminishing switching costs of data. This isn't just about new tools; it's about a fundamental re-evaluation of how businesses create value and maintain competitive advantage. Those who understand this cascading effect--from the commoditization of code to the strategic imperative of owning customer data--will be best positioned to navigate the coming era. This analysis is crucial for founders, investors, and established tech leaders seeking to anticipate and capitalize on the next wave of innovation, offering a strategic lens to identify opportunities where conventional wisdom falters.

The Data Deluge: When Switching Costs Evaporate

The rapid advancement of AI, particularly in code generation, has dramatically lowered the barrier to creating software. This might seem like a boon for innovation, but Siemiatkowski argues it's merely the first domino to fall. The real seismic shift is on the horizon: the commoditization of data switching costs. Historically, proprietary data locked within SaaS platforms created significant inertia, making it difficult and expensive for companies to migrate to new solutions. However, AI agents are poised to automate this data extraction and migration process, effectively dismantling these long-standing moats.

"The next thing that's going to hit everyone bad is the switching cost of data, because so far what you're seeing is you have proprietary data stuck in, for example, the CRM vendor or the other software as a service that you're using currently."

This erosion of switching costs has direct implications for the valuation of traditional SaaS companies. Siemiatkowski suggests that the current multiples for software companies, once stratospheric, may still have significant room to fall. He draws parallels to utility companies, which trade at much lower multiples due to their essential but less dynamic nature. Companies like Chegg, which saw their market value plummet after ChatGPT's emergence, serve as a stark warning. While not every software company will face such an extreme fate, the trend indicates a recalibration of what constitutes sustainable value in a world where software is increasingly easy to replicate.

The "Company in a Box" Economy and the Rise of the Integrated Stack

The ease of software generation, coupled with AI's ability to orchestrate various tools, points towards a future where entire business functions can be bundled into "companies in a box." Siemiatkowski's personal project, "Company in a Box," which combined open-source accounting and CRM software with an AI agent to handle bookkeeping and customer account setup, exemplifies this trend. This model is particularly disruptive for small businesses, where AI agents can now perform tasks previously outsourced to human accountants or customer service representatives.

This leads to a critical question: if AI can facilitate seamless data migration and even replicate entire business functions, why would large companies continue to invest heavily in bespoke SaaS solutions or even internal development for non-core functions? The argument is that as AI agents become more sophisticated, the need to "reinvent the wheel" diminishes. Instead, the focus shifts to assembling and orchestrating pre-existing, standardized components.

However, for companies like Klarna, which view their technology stack as their "operating system," the approach is different. Siemiatkowski explains that for Klarna, integrating AI effectively requires providing it with the richest possible context. This necessitates a unified tech stack where data is not siloed across disparate SaaS applications. This "AI-native" approach, where AI is foundational to the architecture, becomes a strategic imperative for large enterprises aiming to leverage AI for deep operational efficiency and contextual understanding.

"We need to reimagine the tech stack with AI first, being AI native, and incorporate AI and the deterministic and probabilistic code into one tech stack that becomes the operating system of the bank."

This highlights a bifurcation: while many businesses might opt for integrated AI solutions that bundle services, large, tech-first companies may find it more advantageous to build their own AI-centric operating systems to maximize data context and proprietary advantage.

AI in Customer Service: From Cost Center to Competitive Differentiator

The conversation around AI and customer service is fraught with both opportunity and anxiety. Siemiatkowski candidly discusses Klarna's early adoption of AI for customer service, noting that it initially handled simple queries, freeing up human agents for more complex tasks. The surprising outcome was the immediate, significant impact on workload, equivalent to hundreds of agents' worth of work.

This experience underscores a key insight: for AI to be truly effective in customer service, it needs access to deep contextual information, often residing within a company's source code. This realization led Klarna to believe that off-the-shelf solutions might not suffice, as customer service becomes intrinsically linked to the company's core technology stack.

Siemiatkowski also addresses the public perception surrounding AI-driven job displacement. He clarifies that Klarna's workforce reduction was not solely due to AI but was part of a broader strategic shift. He then pivots to a more nuanced view: while AI can handle routine customer service, the future of premium customer experiences will likely involve human connection. Klarna's innovative "Uber model" for customer service, recruiting passionate customers to act as part-time agents, exemplifies this strategy. These customer-agents, deeply familiar with Klarna's products, provide high-NPS interactions, demonstrating a path where AI augments, rather than simply replaces, human roles, creating a unique competitive advantage.

"The future of VIP experience will be the human connection, the relationship."

This approach acknowledges the potential for AI to commoditize basic service while simultaneously creating a differentiated, human-centric offering for those seeking a higher level of engagement.

The Unfolding Future: Data Compression, Enterprise Squeeze, and the New Builder CEO

Siemiatkowski's analogy of AI as a "compression technology" offers a powerful framework for understanding its long-term impact. He posits that AI, by identifying and consolidating redundant information, can drastically reduce the need for compute power in enterprise settings. This contrasts with the generative demands of entertainment or personalized AI experiences, creating a tension between efficiency and creation.

The implication for data centers and chip manufacturers is significant: enterprise demand for raw compute might decrease as AI optimizes data storage and processing. This compression is driven not just by AI's intelligence but by economic realities. Duplicating information and effort is inefficient; reusing existing code and data is not. This mirrors principles seen in platforms like Wikipedia, which enforce strict standards for a single source of truth.

This technological shift necessitates a new breed of leader. Siemiatkowski believes the current era demands that even public company CEOs become builders again, actively engaging with AI tools like Cursor or Claude. The ability to translate ideas into tangible prototypes, even if not production-ready, allows for a deeper understanding of AI's capabilities and limitations. This hands-on approach is crucial for making informed strategic decisions in a rapidly evolving landscape.

"If I meet investors today that haven't actually downloaded and tried to build something themselves, I think they don't have the skill set to make an evaluation of the company they're looking at."

Ultimately, the conversation paints a picture of a future where the software landscape is dramatically reshaped. The commoditization of code and data switching costs will force a re-evaluation of business models and competitive advantages. Those who embrace AI not just as a tool but as a fundamental architectural principle, and who can strategically leverage data and human connection, will be the ones to define the next era.

Key Action Items:

• Immediate Actions (Next 1-3 Months):

  • Assess Data Silos: Map all proprietary data currently locked within SaaS solutions. Identify potential AI-driven tools for data extraction and migration.
  • Experiment with AI Builders: Have key technical leaders and product managers experiment with AI coding assistants (e.g., Cursor, Claude Code) to understand their capabilities and limitations.
  • Pilot AI-Augmented Customer Service: Explore small-scale pilots for AI-powered customer service to handle routine inquiries, freeing up human agents for higher-value interactions.
  • Review Tech Stack Architecture: Begin evaluating the current tech stack for AI-native integration opportunities, prioritizing areas where deep data context is critical.

• Medium-Term Investments (Next 6-18 Months):

  • Develop Data Strategy: Formulate a clear strategy for owning and leveraging customer data, considering how AI can enhance its value and reduce switching costs for your offerings.
  • Invest in Internal AI Expertise: Invest in training and hiring individuals with AI development and integration skills to build proprietary AI capabilities.
  • Explore Integrated "Company in a Box" Solutions: Evaluate off-the-shelf AI solutions for non-core business functions to assess cost savings and efficiency gains.
  • Rethink Customer Service Model: Design a customer service strategy that strategically blends AI efficiency with human connection for differentiated experiences.

• Longer-Term Investments (18+ Months):

  • Build AI-Native Operating System: For core businesses, consider architecting or re-architecting the tech stack to be AI-native, prioritizing deep data context and integration.
  • Develop Proprietary Data Moats: Focus on creating unique data assets and insights that are difficult for competitors to replicate, even with AI-driven data migration.
  • Cultivate Human-Centric Premium Services: Invest in building high-touch, relationship-driven services that AI cannot easily commoditize, creating a distinct competitive advantage.