Time for a Klarna reality check: why it’s not so easy to recreate a transformation.

Time for a Klarna reality check

Last week, financial services company Klarna created a buzz by sharing a massively successful customer service transformation. They now handle two-thirds of their customer service chats with LLM-based AI assistants. The change is delivering amazing outcomes, including a significant profit improvement for Klarna. You might be wondering how you could achieve the same results.

Here’s the catch. Unless you are a large business with a lot of expertise, this transformation will be slow and risky. (That’s where Crescendo comes in … but more on that in a moment.)

The Crescendo team has 25+ years in customer experience and technology. Our experience leads us to believe that though the promise of AI is huge, the path is not exactly easy.

First, we’re talking about something way bigger than just deploying a bot. AI reconstructs and simplifies the technology and staffing of traditional frontline sales and customer service. What’s emerging is a streamlined frontline service with three new areas:

• LLM based AI assistants. These assistants will now handle up to 80% of customer inquiries via messaging and voice. With this change, businesses need to transform their mindset from ‘deflecting’ customers away to engaging them in as many conversations as possible. But beware — AI assistants still require expert setup to avoid embarrassing results.
• Humans-in-the-loop. These are people who step in where a human is preferred, or where an issue needs to escalate beyond them. This next generation of customer service agents also work to continually test and finetune the AI assistants. Other existing frontline agents are moved into higher value proactive or expert service roles.
• AI-based conversation analysis and quality control. AI can now review and quality score every conversation that takes place — whether it’s with an AI assistant, the humans-in-the-loop, or both. This total analysis is essential to ensure successful frontline transformation. Beyond fine tuning the customer experience, it surfaces business insights such as reasons for returns or customer unhappiness. This is how frontline service finally moves from being a cost center to a business catalyst like in Klarna’s case.

The best part of all is that in this new frontline service, we get to finally ditch the terrible things that cost money and make customers angry: IVR/call trees, legacy bots, forms, email, large websites with hard to maintain and localize FAQs, repetitive hold music, post call surveys, and on and on.

How Klarna transformed their service

Klarna got great results, but they had a team of experts working on the problem and were assisted by OpenAI. The to-do list for a company looking to modernize its frontline service is long. Klarna would have had to exhaustively curate and improve its business knowledge to train the bots, integrate with backend systems, set up new workflows, test and tune the AI assistant, and manage the expectations of existing frontline agents.

OK, back to us.

Here’s what we suggest. First, divide your service into frontline and expert. You manage the experts, many of whom are your current agents. Then a company with a specialization in AI and CX steps in to run all your frontline support. That’s Crescendo, and that includes the AI assistant, humans-in-the-loop, and the AI conversation analysis.

Crescendo is your frontline sales and service option for the AI age. We promise to create higher quality, cheaper customer experience. And you won’t need to wait months or get a degree in LLMs. (We have those). We’ll do it all in days — and you’ll only pay for successful outcomes.

Let’s talk about how we can accelerate your transformation.