The role of APIs are changing gradually and surely. Jon Scheele talks to CXpose.tech about the topics he expects will drive conversations.

There is a crucial but often overlooked component that is driving AI innovation forward: Application Programming Interfaces (APIs). As AI agents become increasingly sophisticated, the role of APIs in connecting, enabling, and securing these systems has never been more relevant.

“APIs are a great enabler of AI because these large language models need access to data,” explained Jon Scheele, an experienced technology consultant and producer of the API Days series of conferences in Singapore. “The mechanism for large language models to access data through APIs is quite an important topic, especially given both regulatory and commercial constraints.”

In fact, Jon highlighted that this is one of the key topic areas of conversations during the API Days conference in Singapore, come mid-April.

Emerging challenges in the AI era

Another topic that the conference would likely delve into as well would be the fit-for-use of APIs for modern technologies of today, specifically AI, because “LLMs have to be constantly trained on data.”

Jon Scheele

Jon explained, ”There are some concerns that APIs have been used for several years and are not necessarily fit in their current form for AI agents to easily access the data. So, do APIs need to be redesigned? That’s a big topic also at the moment.”

Will all applications have AI agents integrated into them in the near future?

Jon seemed to think so. “From Microsoft Office to your CRM system, there will be an agent hovering there and waiting to answer your question. This requires software designers to think about how they’re going to make their knowledge base available within the application at the point of time that the user needs it.”

However, this integration brings new security challenges in the form of larger surfaces for compromise. Jon pointed out that security threats are evolving to either access data that LLMs have or to attack and ‘poison’ that data.

Thinking through the API-AI agent dynamic

Looking ahead, Jon identified several crucial considerations for business leaders, particularly around AI agent identity and authorization. “There are some legal aspects that need to be worked out about when an AI agent is performing a task on behalf of an individual or company. What level of authority does it have? Does it have its own identity, or does it inherit the identity of its owner?”

The traditional API onboarding process may need a re-look. “The registration of an API has traditionally been done once, with an onboarding process that companies have applied before they allow a partner to gain access to their API,” Jon explained. “That may not scale very well when AI agents aim to access a lot more data quickly.”

Conclusion

As AI continues to reshape the technology landscape, Jon emphasized that APIs must evolve to meet new demands while maintaining security. The challenge lies in balancing accessibility with control, ensuring that AI agents can efficiently access the data they need while protecting sensitive information.

An organization’s API strategy has to look beyond just connecting systems. It must be able to scale up to support upcoming AI-driven applications seamlessly, securely, and sustainably.

Join API Days Singapore from April 15 to April 16, 2025, as it explores how APIs and AI are reshaping the technology landscape and creating new opportunities for innovation and growth.

Register for this event at this link!