How Will Gen AI Transform API Developer Experience?

Generative AI is reshaping tech, with Gartner predicting that by 2025, 90% of global companies will use it. This growth is driven by AI making complex technologies like APIs more automated and accessible, transforming how developers interact with APIs.

a month ago   •   5 min read

By Bill Doerrfeld
Table of contents

Generative AI is making headlines over and over again.

Gartner, for instance, predicts that by 2025, gen AI will be used by 90% of companies worldwide. This is partially due to AI's ability to make working with complex technologies, such as APIs, more automated and accessible.

Large language models (LLMs) could significantly transform how developers (and end users) interact with APIs, taking low-code/no-code to a new frontier. LLMs could also make understanding documentation easier and quickly spin up sample requests in various languages. In short, the API game is set to get a quick and sudden stimulus shot in the wake of generative AI.

Below, we'll consider some ways in which generative AI will likely enhance the developer experience around integrating with APIs. From initial research to code generation, auto-complete, testing, and more, AI is poised to greatly accelerate the API developer experience. It's hard to predict the outcomes of such a new transformative technology. However, as we'll see, some of the exciting AI-driven experiences mentioned below are already being implemented in production today.

Quick Research and Prototyping

First off, how do you integrate with an awesome API if you don't know how to find it? AI is set to greatly improve the research and discovery phase for new software projects. Since LLMs can suggest the proper architecture and tooling for a specific development project, they are already augmenting how developers find and consume internal and public-facing APIs.

First is leveraging AI to help assess various tools, vendors, and APIs you want to work with.

says Ash Arnwine, Director Of Developer Relations, Nylas, on the Spiceworks blog. He forecasts AI as having many tangible effects on the day-to-day API developer's life. One outcome is dramatically reducing the time and effort required to gather a shortlist of top tooling options for the job at hand.

Supercharging Public API Documentation

Secondly, AI is set to advance how developers understand and interact with API documentation. By training machine learning models on internal developer resources, such as OpenAPI definitions, code samples, and tutorials, API providers can create developer portal assistants that respond with detailed and accurate API integration advice.

One example in practice is Bill, from Plaid, described as "a robot platypus that reads our docs for fun." Embedded into the Plaid API documentation, Bill can respond with instructions on how to program specific API functions given a natural language prompt, including the exact endpoints to call and a thorough sample code.

For example, this is what Bill returns when asked how to retrieve a user's current balance statement:

Bill, Plaid's chatbot

With such advanced code generation capabilities, there is an argument to be made that AI assistants like this could replace the need to reference API documentation in its entirety. In fact, API Strategist and OAI Ambassador Erik Wilde, writing on LinkedIn even goes so far as to suggest AI layers that consume APIs will make current developer experience woes null and void.

We will see a shift in API consumption where DX is not part of the picture anymore because we have AI-based applications consuming APIs. That shift will happen, and with a slightly wider view on API value we can make sure that we are prepared for it.

Generating API Requests Within the IDE

When you think of generative AI in software development, code generation and auto-completion are typically top of mind. Tools like Copilot, CodeWhisperer, and TabNine are transforming how many programmers churn out code, usually leading to improved productivity and satisfaction. One study from Microsoft, for instance, found that a group with access to an AI pair programmer was able to complete its task 55.8% faster than the control group.

Code completion is becoming more commonplace within integrated development environments (IDEs). Therefore, AI-assisted development will naturally involve API integration code. For instance, at the Austin API Summit, Jim Bennett showcased how Copilot can be used to understand and interact with local SDKs within VSCode. While the acceptance rate of LLM-generated auto-complete is still relatively low, it will likely improve as LLMs become more fine-tuned and contextually aware.

Improving API Testing and Observability

Another area AI-driven assistants can accelerate is API testing. Application developers often need to test how to interact with APIs and ensure they perform as expected within their language of choice. That, and they must often test a series of interlinked API flows.

One interesting development in the realm of AI-powered tools is Alfred, an assistant created by Treblle. Named after Batman's iconic butler, Alfred is adept at processing OpenAPI specifications and SDKs to thoroughly understand an API's functionality. Utilizing this deep knowledge, Alfred not only answers integration-related queries but also generates operational code. Moreover, it can craft tests in various programming languages and conduct schema validation. Explore an in-depth exploration of Alfred, including its capabilities and features.

Video overview of Alfred, the AI assistant

Postman similarly has an assistant, PostBot, that can aid in auto-complete for API flows. These sorts of AIs could enhance developer productivity and API experience by significantly speeding up the time to first call and taking the hassle out of routine testing.

Specific testing and development environments could also benefit from using generative AI to create dummy data to query instead of production data, which may contain personally identifiable information (PII). This is especially relevant in highly regulated applications within financial, healthcare, or government environments.

Other Ways AI Will Evolve API Experiences

Modern LLMs are opening up a world where natural-language-driven software development is accessible to seasoned programmers and citizen developers alike. Yet, this new era will hinge on framing developer resources to be correctly ingested by AIs and pruning the training data to avoid hallucination.

Above, we've only scratched the surface of how the current generative AI revolution impacts the API developer experience. Here are some other ways AI might enhance the experiences developers have when using web APIs:

  • Automatically documenting and cataloging APIs to improve tooling discoverability.
  • Extracting meaning from unstructured data returned in API responses
  • Fine-tuning machine-to-machine communication for specific industries.
  • Automating API security, such as generating security policies for APIs or scanning runtime traffic.
  • Augmenting developer dashboards with predictive analysis to forecast usage and costs.

As a result of these advances, AI will likely make APIs increasingly more accessible to non-developers. This is important, given that about half of API users already are non-developers. For more insights into how AI can simplify API documentation, see these AI prompts that help make sense of the most confusing API documentation.

Expanding the Reach and Utility of APIs Through AI

The integration of artificial intelligence into API management and development not only simplifies the technical demands but also broadens the potential user base. By automating complex processes and breaking down data silos, AI enables more intuitive interactions between humans and systems, thereby democratizing access to powerful technological tools. This shift could revolutionize industries by making high-level programming and analysis capabilities available to those without traditional coding expertise.

Furthermore, as AI continues to evolve, we can anticipate even more sophisticated applications that further streamline workflows and enhance security protocols. This includes proactive error handling and real-time adjustments to API operations, which could significantly decrease downtime and improve user satisfaction.

In conclusion, the fusion of AI with API technology holds the promise of transforming the digital landscape. By reducing barriers to entry and increasing the efficiency and security of web APIs, AI is not only catering to the current needs of developers and businesses but is also paving the way for innovative solutions that could reshape how we interact with technology. As we continue to explore and expand these capabilities, the role of APIs will become more central in creating agile, user-friendly digital environments.



Learn how Alfred, Treblle's AI assistant, can transform your API development experience by automating and optimizing your workflows.

Talk to our API Expert!

Spread the word

Keep reading