AI-powered shopping assistant, Klarna is leading the way for financial services when it comes to artificial intelligence (AI) implementation. The company has been able to increase its productivity using the tech without compromising customer experience. So what’s Klarna’s secret and how can firms replicate this level of successful AI integration?
Leon Gauhman, chief strategy officer of digital product consultancy firm, Elsewhen believes that generative AI (genAI) and large language models (LLMs) will have a three-fold business impact for banks: cost saving, redesigning work and, in the longer term, generating new revenue. By combining genAI and LLMs with their wealth of existing in-house, proprietary data and innovative user interfaces (UIs), he maintains that banks and financial services firms can streamline, augment and redesign their business processes.
Klarna clout: four strategic steps for banks and financial services brands to achieve Klarna style-AI enabled transformation
Klarna may be busy preparing for an IPO on the NY stock exchange. But that hasn’t stopped the Buy Now Pay Later (BNPL) specialist pushing ahead with generative AI (genAI) and the technology behind it: large learning models (LLMs).
Sharing the results of over two million conversations completed by its OpenAI-powered virtual assistant, Klarna revealed that its new power tool is already handling two-thirds of all customer service chats across an array of mission-critical interactions from refunds to disputes.
Klarna estimates that its new asset is covering the work of 700 full-time agents – and to the same standards of service. Nine out of 10 Klarna staff now use generative AI including the company’s in-house lawyers.
For financial services brands looking to boost productivity without ruining customer experience, this performance must look like the holy grail of chatbots. Yet it only scratches the surface of genAI and LLMs’ potential. What more could Klarna’s AI assistant achieve, powered by the new capabilities unveiled by OpenAI and Google? Their latest models feature vision as well as voice, text and code. This means the AIs can see the humans interacting with them and their surroundings, and respond appropriately.
Reimagined workflows, cost savings and new revenues form a three-sided business win for companies looking to seize the genAI/LLMs opportunity. However, banks and financial services brands also have the chance to streamline and transform their business processes by combining LLM capabilities with in-house proprietary data.
The latest AI releases allow them to develop highly personalised tools and interfaces capable of closely collaborating with staff. This potentially allows finance players to unlock large amounts of previously untapped creativity and potential.
What does it take to get to that point?
Here are four strategic implementation steps that banks and financial services companies can use to follow Klarna’s AI lead:
Step 1: The groundwork
Finance players with AI ambitions should start by prioritising self-discovery and contextual insights. Having assessed their in-house strengths and weaknesses and developed a clear understanding of the competitive market, the next crucial move is ensuring stakeholders from C-level to operational teams are involved. Together, these cross-silo voices must help design a tailored strategy to enhance productivity and address specific challenges, fostering a sense of collective ownership and contribution.
For example, the European Central Bank (ECB) is currently testing genAI’s ability to speed up basic tasks, including briefing drafts, code writing, data summaries and translation activities. Rather than top management deciding where to apply genAI’s skillset, the ECB canvassed its staff to see where they thought changes might be most effective.
Step 2: Testing the water
Banks and financial service companies should design controlled impact experiments using genAI/LLMs, adopting a test-and-learn approach. “We push everyone to test, test, test and explore,” said Klarna CEO Sebastian Siemiatkowski.
Experimenting will empower staff to leverage proprietary data sets. They will also be able to explore diverse use cases and test the boundaries of what genAI and LLMs can achieve.
This phase is core to generating shared learnings, best practices, and principles for the effective use of GenAI. This could include sharing insights into the capabilities of different LLMs deployed, with Gemini, Llama-3, Claude-3, Stability AI, Mistral, and Command all offering alternative genAI pathways to OpenAI’s tech.
Step 3: Operational roadmap
This stage is about iterating the strategic experiments to expand the scope and scale of a given GenAI strategy. It includes defining key performance indicators, identifying jobs to be done, prioritising specific challenges and objectives that genAI and LLMs will address, and building a service blueprint.
This roadmap should outline the processes, interactions, and workflows that GenAI and LLMs will influence. In doing so, it can act as a visual guide to the technology’s operational layout post-implementation. Deploying a robust, agile, and scalable architecture that extends to everything from data management systems and AI models to integration mechanisms is key to the success of step three.
An example of this stage in action is JP Morgan’s DocLLM, an AI model that offers advanced multimodal capabilities. This includes accurate extraction from visually complex documents. DocLLM has undergone extensive evaluation outperforming other models on a range of known datasets.
Step 4: Risk mitigation
Business value stems from genAI and LLMs being baked into the technological fabric of a given financial service. Inevitably, this demands close integration with data infrastructure, databases and third-party systems – potentially endangering highly sensitive personal or commercial information.
You need to carefully assess and continuously monitor this risk factor. Risk also comes in the form of the major change and uncertainty facing staff involved in AI transformation.
Employees will need to be shown how to operate in a new, supercharged workplace, with an emphasis on a revived relationship with tech including intelligent AI-powered tools that boost productivity and are enjoyable to use.
Ensuring a culture of continuous learning and transparency is critical to ensuring buy-in, providing a safety net of support and guidance.
The contentious issue of job losses comes into play here. However, it’s short-sighted to assume that this will be genAI’s foremost impact. Mastercard, for example, expects its new proprietary AI model, Decision Intelligence Pro, to help banks in its network identify fraudulent transactions in real-time—with cost reductions as high as 20 per cent. This level of saving creates space for innovations, investments, and a whole new calibre of job roles.
Final thought
GenAI and LLMs can help banks and financial services brands reverse entrenched trends around employee experience and job satisfaction. Additionally, they can drive much greater levels of innovation, experimentation and productivity within organisations. Now is the ideal time to perfect the methods, approaches, and skills to create bespoke AI-driven products and solutions.
