Why "AI" can't succeed without APIs
Mega tech trends like the cloud, the mobile phone era, metaverse and now AI all depend on enabling technologies sitting right beneath the surface hidden from nearly everyone’s view. Their structural integrity depends on the flawless operation of those enabling technologies, which in many cases are Application Programming Interfaces (APIs). As such, their success depends on API adoption. Nowhere is this truer than in the rapid proliferation of AI technologies, like generative AI, which require a simple and very easy-to-use interface that gives everyone access to the technology. The secret here is that these AI tools are just thin UIs on top of APIs that connect into the highly complex and intensive work of a large language model (LLM).
It’s important to remember that AI models don’t think for themselves, they only appear to be so that we can interact with them in a familiar way. APIs are essentially acting as translators for AI platforms as they’re relatively straightforward, highly structured and standardized on a technological level. What most people think of as "AI" should be viewed through the lens of an API product; and with that mindset, organizations can best prepare for what potential use cases are possible and how to ensure their workforces have the skills to put them into action.
AI and APIs
Savvy workers are already integrating ChatGPT into their daily office routine, for brainstorming, summarizing long text, as a search engine replacement, for translation or for writing routine emails. But these capabilities are hindered by the fact that it relies on public data available up to 2021, making it difficult to trust in a fast-paced business environment.
All signs are pointing to the development of even more powerful AI models, multimodal AI models (that incorporate text, speech, images) and models that continuously learn from their interactions. These AI models will be used by companies to build AI-backed products for the general market or as in-house proprietary AI applications.Future AI applications will differ from apps like ChatGPT in many ways. The greatest differentiator will be the potential of integrating the company’s data and functionality (accessed via APIs) with an AI model (in turn accessed via APIs). For some tasks, an advanced AI platform might even be able to search for the right API to accomplish the task and engage with it automatically. Many use cases become possible when a company’s data and functionality are available via APIs because they allow an AI model to interact with that proprietary information in a secure and closed manner. For example, some of the easier exercises for an AI model interacting with APIs are probably generating personalized emails for customers or responding to customer service requests. More advanced use cases will involve creating new digital products and touchpoints.
It’s important to note there will be an increased focus on data privacy, security and safety along with the proliferation of AI. While it’s too early to say how this will play out in detail, proper API governance will need to offer established mechanisms to control exactly these aspects: security, dataflow, privacy and access.
Powering the Rise of the Business Technologist
To realize the promised efficiency gains of AI solutions, they must be seamlessly integrated into the day-to-day operations of various business functions. It might be assumed that packaging data and functionality into easy-to-consume APIs is the domain of the IT departments. Traditionally, IT departments were tasked with such integration activities. However, centralized IT departments often lack the business knowledge and contexts to create effective APIs, and the business requirements often change too quickly for their lengthy workloads, leaving them overwhelmed.
Luckily, there’s a solution in the form of business technologists: knowledge workers who are self-sufficient with technical knowledge to perform the relevant integration work much more efficiently. They do not necessarily have the deep-tech skills to write scalable, reliable code that is in accordance with the technical architecture, they can instead utilize low-code developer tools to properly orchestrate and consume APIs. While there are only 27 million developers in the world, there are over 1 billion knowledge workers. It will be easier to upskill knowledge workers into business technologists instead of scouring for increasingly expensive developers. Organizations looking to properly empower them will need to provide thorough documentation, authority to build and manage solutions, and best-in-class APIs to provide an intuitive experience.
Don’t Underestimate API’s Value
At first glance, APIs don’t seem to matter to the AI revolution. AI may seem like magic, but behind the scenes, APIs are essential for requesting results from LLMs and integrating the platforms with existing functionality and data. Investing in API innovation, adoption and usage means investing in the strong foundation essential for all workers that want to ensure that AI will take their companies to new heights.
Subhash Ramachandran is SVP Product Management, Software AG.