Transforming Open Banking Developer Experience with AI

IBM Research describes generative AI as “deep-learning models that can generate high-quality text, images, and other content based on the data they were trained on.” 

In other words, it’s AI that doesn’t just provide simple responses. It can write code, draft tutorials, design UIs, or troubleshoot errors, all by understanding patterns in the data it has seen. For banks, that means AI can become a sort of always-on assistant for developers navigating your APIs. 

This post highlights how generative AI can complement Open Banking APIs, improving developer experience, onboarding, support, and prototyping. 

Here’s how it plays out in practice, with some examples using OBP’s Opey!

Open Banking has already made huge strides in improving developer experience. Gone are the days of screen scraping and juggling 1:1 connections. Still, onboarding developers can be pretty slow and frustrating.

Developer experience is the overall ease with which a developer ecosystem uses your API. Better experience leads to less bottlenecks, faster builds, and an overall healthier ecosystem. 

Open Banking introduced interfaces that transformed most of this experience to satisfy today’s technological expectations. However, banks in Europe spent over 5 years implementing a satisfying and functional API portal. 

A chatbot with knowledge about your API can provide tailored getting started guides for each developer, guiding them from registration all the way to their first successful API call.

Below is an example from the upcoming version of our developer portal, with an agentic twist:

While the user can ask direct questions, the portal provides template prompts for the most frequently asked questions to help the developer kick off the tutorial conversation, namely Getting Started, Authentication, Consent, and SDKs. 

As banks evolve their API stacks and become more technology-driven, their catalogues will grow and so will the complexity – even for internal teams. Authentication flows, required parameters and endpoint bundles and combinations can become overwhelming. 

Even experienced developers can spend significant time digging through documentation to find the right combination for a given use case.

AI can streamline this process by recommending the correct endpoints, guiding developers through authentication, and even provide example requests or scripts. This reduces delays, lowers costs, and keeps developers engaged.

Not all banks have a dedicated API team. Smaller teams often struggle to keep up with Fintech developers testing and integrating APIs. 

Enter the always-available assistant to reduce the burden on technical support teams. Traditional chatbots quickly hit a wall and have to refer users to human beings – especially when faced with complex or nuanced questions. AI can troubleshoot errors, point out syntax issues, and provide the correct format as a code sample.

One of Open Banking’s biggest advantages is fast prototyping — sandbox environments and accessible APIs let developers test new ideas quickly. 

Generative AI amplifies this, streamlining prototyping by allowing developers to generate functional code snippets, build API workflows and simulate interactions within seconds. Some developers might request some CSS code snippets for a proof of concept, beginners might ask the chatbot to provide examples for a specific use case they’d like to start.

For instance:

• Automating PoC Development: A developer could ask, “How do I retrieve account balances and recent transactions?” and get not only the right endpoints but also an auto-generated script in Python or JavaScript.
• Instant UI Mocks: Some developers may request AI-generated CSS or front-end code snippets to create proof-of-concept interfaces that visualise Open Banking data.
• Low-Code/No-Code Assistance: AI can enable non-developers to prototype financial products by guiding them through assembling API interactions without deep technical expertise.

Up to now, we’ve been talking about AI giving suggestions. But we’re already in the era of agentic AI — AI that can act on your behalf, with consent.

Instead of generating a guide on how to create a bank and related accounts, why not let the agent do it for you? This won’t only facilitate actions for developers, but it will also lower barriers for non-technical users who want to build applications or test workflows. 

Vibe coding is quickly becoming a way for anyone with an idea to be able to build it, whether they’re a programmer or not. For banks, this opens possibilities for internal teams too — marketing, product, or analytics teams can interact with APIs without writing a single line of code.

The push for better banking interfaces didn’t happen in a vacuum. One of the reasons it became necessary was the surrounding market providing innovative applications. Users quickly got used to multi-channel access, biometric logins, and even small UX details like rounded corners.

A significant portion of internet users are starting to expect chatbots. Developers and non-technical users now expect these chatbots to provide more than links — they want tailored, interactive guidance.

The use cases we mentioned aren’t exactly necessary, but user expectations don’t always follow what is “necessary”. Meeting these evolving expectations can improve engagement and overall make your API offering feel modern and approachable.

The market is going a bit wild, putting AI everywhere and anywhere it will fit. However, AI is best implemented pragmatically, implementing it where it will make a difference. Open Banking has already shown how improved interfaces can benefit both developers and customers.

Generative AI offers banks the chance to elevate this experience further — streamlining onboarding, guiding developers through endpoints, enhancing technical support, and enabling rapid prototyping. Even if your team isn’t ready to go full AI agent yet, starting small can help keep pace with expectations and start offering a more modern developer experience.