Klarna CEO Predicts AI Will Kill SaaS
Klarna CEO Sebastian Siemiatkowski argued that AI agents will fundamentally undermine traditional SaaS business models by making it nearly frictionless for customers to switch between platforms. He predicts this collapse in switching costs could cause software company revenue multiples to fall from 20-30x to as low as 1-2x. Siemiatkowski also noted that AI has already allowed Klarna to reduce its workforce from 7,000 to under 3,000.
- Klarna's AI assistant, built with OpenAI's technology, now handles two-thirds of all customer service chats, equivalent to the work of 700 full-time agents. This has resulted in a 25% drop in repeat inquiries and has shortened the resolution time for errands from 11 minutes to under 2 minutes. - While the company projects a $40 million profit improvement in 2024 due to AI efficiency, it has also faced challenges. After an initial push towards automation led to a drop in customer satisfaction, Klarna began hiring human agents again to handle complex issues, acknowledging that its cost-cutting focus had negatively impacted the customer experience. - The rise of AI agents is attracting significant venture capital, with investors pouring billions into startups that create autonomous AI systems. In the first half of 2025 alone, agentic AI startups raised $2.8 billion in global funding. - Top VC firms like Andreessen Horowitz, Sequoia Capital, and Founders Fund are actively investing in AI agent startups. Notable examples include Sierra, a customer service AI agent startup founded by the former co-CEO of Salesforce, which has reached a $10 billion valuation. - In the real estate sector, AI is being used to automate property valuations, generate leads, and create virtual property tours. Companies like Zillow use AI to analyze images and property data to provide "Zestimates," while platforms like Homebot use AI to identify homeowners who are likely to transact soon. - The fitness industry is leveraging AI to create personalized training and nutrition plans based on real-time data from wearables. This technology can analyze metrics like heart rate and sleep patterns to optimize workouts, predict injury risks, and tailor recovery protocols for athletes. - The core argument against traditional SaaS is that AI agents act as an "intelligence layer" that can interact with multiple software backends via APIs. This could commoditize the underlying software, shifting customer loyalty from a specific SaaS platform to the AI agent that delivers the best outcome, regardless of the software it uses. - This shift is prompting a move away from per-seat SaaS licensing toward usage-based or outcome-based pricing models. The value is no longer in access to the software itself, but in the results the AI agent achieves using that software.
