API Trends for 2024: A Glimpse into the Future of Interface Technologies
In 2023, the surge of Generative AI captured most of the spotlight in the tech world. I believe it marked an inflection point not just in how we interact with machines but also in society as a whole. As we move further into 2024, I believe that the API landscape is set to undergo a significant and profound evolution, I believe largely driven by the need to support a diverse array of new use cases emerging from what I like to call the AI Industrial Revolution.
This post delves into the top trends in the API landscape, combining insights from my own experiences as an API product executive and practitioner, along with valuable perspectives by other API experts and thought leaders featured in the API Futures collection.
Without further due, following my predictions for 2024:
1) The Rise of Polymorphic Interfaces for AI
As Kristof van Tomme said on his article APIs are interface utilities:
"As LLM AI systems start consuming APIs, this evolutionary block might get lifted. When AI-driven API consumption starts to reduce the value of applications and increase the importance of APIs even more, the art of handcrafted API design will come under pressure"
Couldn't agree more. It is my believe that the forefront of API evolution will be marked by the emergence of polymorphic interfaces tailored towards AI integrations. Unlike traditional APIs (and best practices around it), these interfaces cater specifically to AI's unique consumption patterns, embracing unconventional design principles. For instance, a field named "mandatoryPhoneNumber" might defy standard API design principles (and perhaps even irritate API purists) but proves very effective in AI integrations to prevent a model to make up a phone number (a hallucination) rather than expecting the user to enter it.
There are many other examples which I won't be listing in this post, however I hope the point is clear and we're only seeing the tip of the iceberg. I don't think anyone knows for sure which direction will this take (don't think even the OpenAI board nows judging from the 2023 dramas!) but my guess is that as large enterprise settle on their GenAI strategies (in 2023 majority were just still getting their heads around it) there will be a surge on specific AI-led integration needs driven by GenAI adoption. Startups like PolyAPI.io and Superface see such opportunities and are positioning well to ride this new wave.
2) Polyglot APIs are here to stay
While REST continues to reign supreme in the API kingdom, other array of protocols are also gaining prominence (e.g. GraphQL, gRPC, Websocket, tRPC, etc) and I believe such trend will only intensify in 2024 and beyond.
It is my opinion that a polyglot environment enriches the API ecosystem, offering more tailored solutions for varying needs and applications. It signifies a move away from a one-size-fits-all approach, acknowledging the myriad ways in which APIs are now consumed and implemented. Of course, this isn't news; Microservice Architectures have, since their inception, embraced the adoption of polyglot programming languages and technologies to build services. However, amongst API practitioners, there was still a fair amount of debate on this. I believe that this year the dust will settle and there will be a majority recognising that there is room for more than one ring! (some will argue that such recognition has already happened!).
Companies like Postman and many other API management vendors clearly recognise this trend and have already incorporated support for a wide variety of protocols in their products. Startups like PolyAPI (previously mentioned) provides such support too.
Few articles of the API Futures collection also touch upon this point, some that come to mind:API trends 2024 by Daniel Kocot and API Sprawl to Be a Pressing Concern in 2024 by Bill Doerrfeld.
3) Broad adoption of API Product Management with equal focus on API consumption
One point that was highlighted by many experts in the API Futures collection is the believe that API Product Management as a distinct discipline will intensify this year. This trend reflects on the growing complexity and strategic importance of APIs in business ecosystems.
As a someone responsible/accountable for an API Product and overall API strategy/governance across a business unit, I welcome this and also echo it. In fact I wrote about this in my previous blog What is API Product Management.
API Product Managers are pivotal in bridging the gap between technical API development and business strategy, ensuring that APIs not only function well but also align with broader organisational goals. In other words, producing Good APIs that customers want (or better, need!).
In practice however, this is easier said than done. API Product Managers are required to have more than just technical know-how; they must also possess deep expertise in their respective business domains and also the business itself. I firmly believe that the criteria for what constitutes a 'good' API can vary greatly across different industries and specific use cases. Sure, there are common denominators, but the level of specialisation needed to tailor API products to meet unique market demands must not be underestimated and overlooked.
Lastly, in 2023 I observed a growing popularity for SDKs. Although SDKs are not new (far from it!), it feels that the expectation is shifting towards all good APIs having to also provide good SDKs. This belief is reinforced by the emergence of new players in this space, such as Speakeasy, Stainless, and the previously mentioned PolyAPI and Superface, who also play in this area. As a result, I believe that the scope of API Product Management will continue to evolve, increasingly focusing on the consumption side of things.
4) Others worth mentioning
- API workflows: I am quite excited about the emerging API Workflows specification, perhaps because I have a very clear use case for it—which at present we're satisfying by using Postman Collections (see an example here). However as collections aren't really meant for this (at least not right now), many details (e.g. variables passed across interactions, rules to bear in mind, etc) are completely omitted. I think the workflow spec is filling a genuine gap, and I believe that throughout 2024 we'll see increased adoption of it. In fact, I also believe that the workflow spec can play a significant role in defining AI interfaces, as currently there isn't a clear way to communicate a business sequence to AI. Sure, LLMs like GPT-4 are very clever at determining what the next function call could be, but such integrations could be simplified by being able to better communicate the desired workflow to the machine. I recommend Frank Kilcommins' article The API Future is Bright with the New API Workflows Specification for more details on this.
- Intensified API security: As APIs continue to become more integral to business operations and ecosystems, there will also be an increased focus on API security. With the proliferation of APIs in all of its flavours / styles, the risks associated with data breaches and unauthorised access also will rise. This challenge is getting intensified by bad guys turning to Gen AI to craft sophisticated attack vectors. Consequently, there is an undeniable need to proportionately prioritise API security measures and I think we will see just this in 2024. I recommend the following articles for more insight on what to expect in 2024 regarding API security: Protecting against API Token and Credential Theft in 2024 by Gary Archer and What API security may look like in 2024 by Dana Epp.
Conclusion
2024 is turning out to be a truly interesting year for anyone working with integrations and APIs. I hope it's evident to most of us that we stand at a pivotal evolution point in technology, heavily influenced by Generative AI and the new API usage patterns that it will drive. Furthermore, this evolution is also fuelled by other trends, such as Polyglot APIs, which in my opinion are here to stay, emerging specifications like API workflows, and an increased need for security to protect against the next generation of LLM-powered attack vectors. All these factors highlight the need for skilled product managers who can bridge the gap between technology and business needs, thus ensuring that value is delivered to both the business and its customers.
For more predictions I recommend you read other articles in the API Futures collection.
Comments
Post a Comment