The Center for Data Innovation spoke with Sagar Batchu, co-founder and CEO of Speakeasy, a company that provides an up-to-date platform for building application programming interfaces (APIs). Batchu discussed how Speakeasy’s platform accelerates API development, the role of APIs in data innovation, and how AI is shaping the future of API design.
Martin Makaryan: What does Speakeasy do?
Sagar Batchu: Speakeasy helps developers build APIs more easily and efficiently. Most companies today have data they want to expose securely and consistently to their customers or need to interface with data from other sources. APIs are the core technology that powers these integrations. Speakeasy fits neatly into a company’s existing software architecture, providing building blocks that help end users more effectively access their digital product. We generate up-to-date, ready-made examples and tools like software development kits (SDKs) in various programming languages, which gives the developers flexibility and ease in working with APIs.
Makaryan: How do APIs enable data innovation?
Batchu: Good APIs are crucial for data innovation because they provide a safe, simple, and constrained way of exposing data. If you look back 20 years ago, companies would expose entire databases or provide file exports, which was inefficient and potentially risky. Today, APIs allow data to be sent in multiple ways: on-demand, streamed, or in batches. This flexibility has enabled innovations like ChatGPT and other AI experiences we are familiar with as consumers. Underneath, the end-user applications connect through APIs. For instance, ChatGPT uses an API that streams back words or tokens from the AI model. This standardized protocol for streaming data from AI models is an exciting innovation in the API space over the last couple of years.
Makaryan: How do you use AI?
Batchu: We use AI to streamline and automate the API design process, in turn empowering our customers to deliver data innovation through engaging APIs. API design is iterative and involves writing technical specifications that conform to standards like OpenAPI. We use generative AI to guide customers through building and validating these documents. We also offer a chatbot that can provide feedback and help customers make good decisions on API design by constraining the open-ended nature of the process.
Makaryan: What challenges have you faced as an innovator?
Batchu: As a first-time founder coming from a developer background, one of the biggest challenges has been shifting from pure development to finding product-market fit and working directly with customers. In established companies, developers are often secluded from customer interactions. In a startup, especially at our stage, we have had to break down those boundaries. Everyone in our company, regardless of their primary role, works directly with customers. This direct interaction has been a significant shift from my previous experience and while rewarding, it has also been a challenge.
Makaryan: Looking ahead, what trends do you anticipate will impact Speakeasy?
Batchu: We are seeing exciting developments in the AI space, particularly in what we call agentic interactions. Large language models (LLMs) are becoming more powerful, not just in chat-based experiences but in taking on workflows and carrying out tasks. For example, an LLM-powered assistant could book a plane ticket by making a series of API calls under the hood. This development is highly relevant to us because it requires both existing tools and new ones. For instance, an AI model does not need a full documentation site to access an API: it needs an accurate description of the input data structure required and the expected output data. We are well-positioned to offer companies the tools to make their APIs accessible to LLMs. While we are still in the early stages of this evolution, we have built up trust with our customers to be the go-to provider for these tools as the technology evolves.
