🤔 AI in Banking: Context Matters

I’ve been thinking a lot about AI and the consequence for banking a lot lately. (And I can assure you that this is not yet another article where someone has asked an AI what it can be used to) There has been written a lot about jobs disappearing or being completely altered in the years to come because of AI. Bill Gates as timely stated that “The Age of AI has begun”, but what are the consequences for how users interact with their banks?

The easiest first conclusion people jump to is to get personalized financial advice from an AI. Imagine the AI analyzing all your spending and saving patterns and providing you with personalized financial advice. This is the holy grail for banks, but likely won’t happen in the short term. Why? The underlying data isn’t good enough. Yes, you’ve used 10k at Rema 1 000 the last month, but what did you use it on? Yes, you spent thousands on a Spa last month, but what if it was a gift for someone? Similarly, it needs to factor in details such as electricity pricing, type of house, and whether to accurately provide advice on your electricity bills.

Context matters when it comes to personalized financial advice, and that’s where AI in banking has a long way to go. Simply analyzing spending patterns is insufficient to provide valuable feedback, especially if credit cards from elsewhere are involved. But in other aspects, I belive AI will make banking more efficient, especially when it comes to answering simple customer inquiries or automating fraud detection.

Another conclusion that is easy to jump to is that AI can have a huge impact on credit risk assessments. After all, what machine learning is good at is analyzing vast amounts of data and categorizing similar patterns. But I can assure you that you wouldn’t have a banking license for long if your documentation was that you have a black box that sorts out the best customers. There is, however, merit in using AI to optimize business rules and threshold limits based on data. Making sure you are structuring your business rules with decision matrixes or DMN tables is the first step in making sure you’re positioned to optimize your credit assessments with AI.

🤖 Stripe, Klarna, and Shop join ChatGPT's plugin store

You’ve been sleeping under a rock if you haven’t heard about GhatGPT by now, but last week they made an entry into Fintech by announcing plugins. Three of their first plugins: Klarna, Stripe, and Shop, Shopify’s consumer app.

Consumers can add the plugins from ChatGPT's plugin store (ChatGPT Pro). All they have to do to activate Klarnas app, as an example, is to ask ChatGPT for shopping ideas to be presented with tailored products. ChatGPT will decide when to use the plugin based on the conversation. Shoppers can give further instructions or ask for more product recommendations. By clicking on the product link, they will be taken directly to Klarna's search and compare tool to compare prices across different brands.

Klarna announced their plugin with this statement from Sebastian Siemiatkowski, Co-founder and CEO of Klarna: 

Well, I might disagree with Sebastian about whether a ChatGTP plugin (you have to be a paying subscriber for) passes the “mom test” and whether my mother gains from being able to purchase more items easily. But it is an interesting first iteration!

📈 5 trends shaping the future of finance

Itera has launched a report about 5 trends shaping the future of finance.

These are trends that will present both challenges and opportunities, shaping innovation and strategies in the years to come. The report's goal is not to solve all the challenges faced by the industry but rather to create a basis for innovation, development, and discussion of the future. Not surprisingly, one of the key trends is responsible AI. The rest of the key trends are:

1. Fluid expectations: Great user experiences from other sectors create high expectations for financial services.

2. Integrated finance: Financial actors can achieve profitability by taking a back seat.

3. Sustainability and integrity: The sustainability trend continues, and the financial industry is responsible for leading the way.

4. Future anxiety: High prices, war, and pandemic aftermath affect consumers.

5. Responsible AI: The possibilities seem endless, but missteps can be costly.

🥴 Why is it so hard to do great work at scale?

George Kedenburg, product design lead at Humane and previously at Instagram and Facebook, has written about why it is hard to create great work at scale. He has summed his theory into this formula:

He also lists a lot of suggestions for how you can balance this out, f.ex. by introducing a craft or systems-focused counterbalance to metric goals so that success doesn't come at the cost of user experience. If you’re working on a product, I highly recommend reading the article!