In his letter to shareholders accompanying Klarna’s Interim Results released in August 2024, CEO Sebastian Siemiatkowski made a bold announcement that left many in the industry stunned. The Swedish fintech company, which had already been an early adopter of artificial intelligence (AI), was set to reduce its workforce by nearly 50%, going from 3,800 to around 2,000 employees. This wasn't a result of poor financial performance but rather a deliberate strategy to use AI to fundamentally transform the business. Siemiatkowski, ever the advocate for technological advancement, explained that Klarna could now “do much more with less.” The company’s AI-powered systems, particularly its customer service assistant, were doing the work of 700 full-time employees.
Klarna’s use of AI is not just about cutting costs. Earlier in 2024 they had revealed how their AI assistant, powered by the same technology as OpenAI’s ChatGPT, was not only more efficient but also delivering higher customer satisfaction. Within its first month of deployment, Klarna’s AI assistant had handled two-thirds of customer service inquiries, reducing resolution times from 11 minutes to just 2 minutes. This impressive efficiency boost translated into an estimated $40 million in profit improvements for the company. Klarna’s journey reflects a broader trend in how AI is revolutionising work, but it also raises essential questions about what this means for employees, industries, and the future of business.
The potential benefits of AI agents in the workplace are profound. Firstly, operational efficiency skyrockets when AI agents handle repetitive, mundane tasks. Tasks like customer service inquiries, data entry, or appointment scheduling can be completed in a fraction
of the time it would take a human. This frees up human employees to focus on more strategic, creative, and impactful work.
Secondly, AI agents bring cost savings. As seen in Klarna’s case, the company managed to significantly reduce its workforce while increasing revenue per employee. This trend is not unique to Klarna; many organisations are looking to AI to optimise operations and cut
staffing costs. Research by McKinsey found that when using generative AI in the customer service environment, issue resolution increased by 14 percent an hour, while time spent handling issues went down 9 percent.
Another key advantage is the improvement in customer service. AI agents can provide 24/7 support in multiple languages, ensuring that no query goes unanswered. For example, Klarna’s AI assistant supports 35 languages and operates in 23 markets, offering global customer service coverage at a speed unmatched by human agents. AI agents can also continuously learn from interactions, refining their responses and decision-making over time.
While the benefits are clear, the path to fully adopting AI agents is not without its obstacles. Many organisations, particularly in regulated industries, face significant trust and compliance issues. A solicitor at one of our legal clients commented recently, “Right now, there’s no way the SRA (Solicitors Regulation Authority) and our professional indemnity insurers are going to let us automate reporting to clients.” This sentiment highlights the concern many have about automating sensitive processes where errors could lead to legal liabilities.
Regulatory bodies and professional organisations often impose strict guidelines that make the automation of high-stakes tasks difficult, if not impossible, without extensive oversight. These concerns are particularly prevalent in sectors like finance, healthcare, and law,
where human judgment is considered essential to mitigate risks. Trust is also an issue with internal stakeholders; employees may fear that AI will render their roles obsolete, leading to job insecurity and resistance to AI adoption.
Furthermore, current AI systems, though advanced, are not infallible. Klarna’s success is notable, but not all companies have the same level of confidence in AI agents’ ability to deliver consistently reliable results. AI “hallucinations,” where systems generate incorrect
or misleading information, remain a significant concern, especially in mission-critical applications.
So, how do companies overcome these hurdles? The solution lies in a gradual transition from AI assistants to more autonomous AI agents. Many companies, rather than diving headfirst into full automation, are starting with tools like Microsoft’s Copilot or Google's Gemini in a user augmentation role. These AI assistants help organisations speed up simple tasks while maintaining human oversight. Microsoft’s bet is this will lead customers to take the next natural step in using their Copilot Studio, part of its broader push with the Copilot platform, which allows businesses to create customised AI agents that can handle specific tasks, such as data retrieval or task execution, within a controlled environment.
The key to success is building trust—both with employees and within the broader regulatory and business ecosystem. Initially, companies can deploy AI to handle lower-risk, repetitive tasks, allowing employees to verify AI output and gain confidence in the technology. For example, Microsoft’s new Copilot Pages feature enables AI-generated content to be iterated on by teams, ensuring that the AI acts as an enabler rather than a replacement. Over time, as AI systems demonstrate reliability, organisations can automate more complex workflows.
In regulated industries, compliance frameworks must evolve in parallel with AI technology. Companies will need to work closely with regulators to ensure that AI systems meet ethical and legal standards, particularly in tasks involving sensitive data or decision-making.
The story of Klarna is a compelling example of the transformative power of AI, but it also serves as a reminder of the complexities involved in fully integrating AI into business operations. The benefits of AI agents—enhanced efficiency, cost savings, and improved
customer service—are clear, but they come with challenges related to trust, regulation, and the human workforce.
Organisations that wish to embrace AI must do so thoughtfully, starting with smaller implementations and gradually building towards more autonomous systems. With careful planning, robust governance, and a focus on building trust, the future of work with AI agents promises to be not just efficient but transformative for both businesses and employees alike.