The Business of AI: The API Advantage—Getting to Work with Generative AI

Sales and marketing leader Brian Silverman outlines the ways APIs act as the wiring that enable AI solutions by connecting data and applications

Brian Silverman February 10, 2025

This is the fifth article in the series, “The Business of AI.” In case you missed them, here are the first four articles:

1. Focus on ‘The Business of AI’ to Move From Hype to ROI

2. A Look at the Different Types of AI and Their Value

3. From Foundation Models to Large Language Models

4. Current Applications and Data Fuel AI Innovation and Solutions

If data and applications are the fuel that trains and powers AI, APIs are the wiring that connects those elements to make innovative AI solutions possible. It’s what enables AI applications to move beyond simple Q&A chatbots, to powerful solutions that complete tasks, integrate into workflows and drive innovation.

In fact, APIs are so important to AI implementation that Gartner predicts “by 2026, more than 80% of enterprises will have used generative AI APIs or models, and/or deployed GenAI-enabled applications in production environments, up from less than 5% in 2023.”

How APIs Enable AI

AI solutions, such as agents, communicate and collaborate with other AI systems.
AI applications call services to complete transactions, schedule service calls and automate processes.
AI systems exchange data with databases, applications and knowledge repositories.
AI integrates seamlessly into complex workflows, enhancing productivity and decision-making.

Without APIs, AI solutions cannot access the data and applications necessary to deliver real business value.

AI APIs in Action

Financial Services – Ntropy’s API standardizes unstructured financial data from various sources such as bank statements and receipts, enabling AI-driven financial applications to interpret transactions with human-like accuracy in milliseconds.
Healthcare – Corti’s AI-powered API analyzes patient interactions in real time, providing medical professionals with decision support and reducing diagnostic errors.
Manufacturing & Supply Chain – o9 Solutions uses AI-driven analytics and APIs to ingest real-time supply chain data, enabling dynamic demand forecasting and inventory optimization.
Retail & Ecommerce – Constructor leverages machine learning and APIs to optimize eCommerce personalization through tailored recommendations and dynamic search results.

API-Enabled AI Customer Service Example

A consumer appliance manufacturer faces an overburdened customer support team. They implement an AI-powered customer support chatbot that can handle routine queries, freeing up human agents for complex cases.

To achieve this, they:

Train an AI model using APIs to access product data, customer service call logs and relevant documentation
Integrate the AI chatbot with ERP, CRM, service scheduling tools and unstructured data sources via APIs
The chatbot leverages APIs to retrieve real-time product manuals, service notices and support documentation
AI logs interactions, updates the CRM and schedules service appointments through API calls.

For a deeper dive into API standards and best practices for AI-driven solutions, see the API Standards Addendum, which outlines key API types, their use cases, and considerations for legacy system integration.

Managing APIs for AI: Governance, Security and Cost Control

To integrate AI APIs effectively, organizations must align API strategies with AI governance, security and cost management.

Governance Considerations

Ensure API policies align with organizational compliance and security standards.
Establish governance for AI agents accessing APIs, defining rules for data access and digital agents’ rights.
Organizations must establish governance policies that define how AI APIs and AI agents access, process and interact with data, ensuring compliance with security and business requirements. Unlike human users, whose access privileges come with an understanding of responsibility and discretion, AI operates strictly within programmed parameters. Therefore, policies must account for how AI APIs and agents interact with each other as well as the human workers who may have different access rights. This ensures clear rules on when AI can act independently, when human oversight is required, and how discrepancies in access levels are managed to prevent unintended data exposure or decision-making conflicts.

Security & Risk Management

AI introduces new risks as APIs become even more valued attack vectors; implement strict access controls and authentication.
Secure AI-enabled APIs to prevent unauthorized data access and ensure compliance with cybersecurity policies.
Identity and access management for AI APIs – AI APIs, whether used by AI agents or other applications, must have clearly defined and enforced access privileges. Implement role-based access control (RBAC) or attribute-based access control (ABAC) to align AI permissions with governance policies, ensuring that AI systems operate within their intended access boundaries.

API Standards and Documentation

Implement standardized documentation practices to ensure transparent and efficient API usage.
Maintain an API inventory to track dependencies between AI solutions and business applications.

Legacy Integration

Identify critical legacy systems that lack APIs and develop a plan to develop appropriate middleware solutions for API connectivity.
Ensure seamless API integration for hybrid AI and traditional IT infrastructures.

Controlling AI API Costs With API Management

AI-driven API usage can become costly due to unpredictable token consumption. Effective API management solutions can help with:

Usage monitoring – Track and control AI API requests and usage patterns.
Rate limiting and quotas – Set consumption caps to prevent excessive API usage.
AI gateways – Monitor AI models and token consumption using API analytics tools.
Prompt optimization – Reduce token use by refining prompts and limiting inference time.

Connecting AI to ROI

A successful AI implementation is not just about choosing the right models—it’s about ensuring that APIs effectively connect AI with the broader business ecosystem to drive long-term value and return on investment (ROI). To achieve this, organizations must regularly audit their API ecosystem, integrate AI API management into their broader IT strategy and proactively address governance, security and cost challenges.

After all, AI depends on data and applications to fuel innovation, but without a well-structured API strategy, businesses cannot fully realize AI’s potential.

Addendum: API Documentation and Standardization

Well-defined API documentation and specifications are critical to the success of AI-driven API solutions. The OpenAPI Specification (Swagger) helps ensure APIs are well-structured, discoverable, and easy to integrate.

OpenAPI (Swagger) Specification

OpenAPI is a widely adopted standard for defining APIs, improving interoperability, and documentation.

Applicable to: Representational State Transfer (REST), Graph Query Language (GraphQL), Webhooks and some Streaming APIs
Limited for: Google Remote Procedure Call (gRPC) (can use gRPC-gateway to generate OpenAPI specs)
Best for: Organizations looking to streamline API onboarding and ensure consistency across AI-driven solutions
Example: AI solution providers publishing APIs with OpenAPI documentation for easier integration by developers

Considerations for Legacy Applications

Many enterprises still rely on legacy applications that use older API standards, such as SOAP or XML-RPC. Modernizing these systems for AI requires careful planning:

Legacy APIs (SOAP/XML-RPC): Older systems may require API gateways or middleware to translate SOAP/XML to modern API standards like REST or GraphQL.
API Gateways: Platforms like Apigee, AWS API Gateway and Kong help bridge legacy systems with AI solutions by providing security, monitoring and version control.
Middleware for AI Integration: Tools like MuleSoft and IBM App Connect facilitate seamless AI integration into legacy ERP/CRM systems.

Choosing the Right API for Your AI Solution

Selecting the best API type depends on several factors:

API Type	Best For	Example Use Case
REST API	General-purpose AI integrations	AI-powered chatbot fetching product details
GraphQL	Reducing data over-fetching	AI-powered analytics tool with complex queries
gRPC	High-performance, real-time AI processing	AI fraud detection with real-time transaction streaming
Streaming API	Continuous AI-driven data analysis	AI-powered market sentiment monitoring
Webhooks	Event-driven AI workflows	AI-driven inventory alerts
Hypermedia as the Engine of Application State (HATEOAS)	Dynamic AI automation and interactions	AI-driven workflow automation
OpenAPI Specification (OpenAPI)	Standardized API documentation	AI solutions offered via external APIs

A Standardized, Deliberate Approach

Maximizing the value of API-driven AI solutions requires a standardized approach, including strong API management and governance.

By carefully selecting API standards and leveraging API gateways or middleware where needed, businesses can create scalable, flexible and secure AI ecosystems that drive long-term success.