What does your A(P)I strategy look like? 

“Tell me what your API strategy looks like, and I’ll tell you how your AI initiatives are going.”

That’s a simplification, of course, but there’s a hard kernel of truth in it. APIs (Application Programming Interfaces) are nothing new; the concept has been around for decades. Yet the ability to reach applications and share information through API-based integration is once again taking centre stage. The reason is simple: the single most important success factor for AI is access to data, and the quality of that data.

Architecture, technology, compute and AI models all play their part. But data quality, and what makes it possible in the first place (being able to reach and consolidate data from different applications), is the biggest single factor for both model development and reliable outcomes. Gartner predicts that organisations will abandon 60 percent of AI projects that lack AI-ready data.¹ ²

With agentic AI, where a language model is combined with tools and goals, this becomes even clearer. An agent that is meant to act in your business has to be able to understand and use the systems you already have. The Model Context Protocol (MCP) has quickly become the open standard for giving agents access to those tools, and in practice it is usually existing APIs that get exposed through MCP. That also means API documentation becomes functional: an agent will only use what it understands.

This is why the API platform and its gateway become a control point for AI as well. That’s where you govern who, human or agent, gets to see and do what. That’s where you protect back-end systems from overload and trace what is actually happening. And that’s where you make information available to the whole organisation, and to partners when you choose to.

Three questions to ask yourself:

A solid API and integration strategy, in other words, lays the foundation for reliable information supply, and it’s the starting point for AI initiatives that deliver real results.

The choice of platform matters too. For the tenth year in a row, IBM was named a Leader in the 2025 Gartner Magic Quadrant for API Management, and ranked first in five of six use cases in the accompanying Critical Capabilities report. API Connect is now part of IBM webMethods Hybrid Integration.³

At epical, we’ve helped companies and organisations across the Nordics with integration and API solutions for more than 25 years, and we’ve worked closely with IBM since 2001, today as a Platinum Partner.

We cover the fundamentals of modern, effective integration in The IT manager’s guide to future-proofing systems integration, which you can download from our Insights page. Want our 10-step checklist for your first API project? Send us a message with the subject line “API guide” and we’ll email it to you.

Take a deeper dive

¹  https://www.dwarkesh.com/p/pretraining-progress-is-mostly-data

 ² https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk

³ Gartner, Magic Quadrant for API Management, 1 October 2026.