2 min read

Simplifying use of Custom Controls in IBM Cognos Analytics with AI

Custom Controls in IBM Cognos Analytics

Over ten years ago, IBM introduced custom controls in Cognos Analytics 11.0.4. The concept was powerful and a framework that lets developers inject JavaScript and JSON configuration directly into a Cognos report, extending what the platform could do beyond anything IBM shipped natively.

IBM created the Custom Control API on top of their robust Enterprise Reporting System and simply said, "Here is the infrastructure. Imagine and build what you need."

The community noticed. The interest was real. But the adoption was not. WHY?

The reason was simple: building with the custom control framework required deep JavaScript expertise, an intimate understanding of how Cognos renders its DOM (Document Object Model) at runtime, and the patience to navigate documentation that was scattered across IBM's developer resources and community forums. For most organizations:

  • The development time was prohibitive.
  • The expertise was expensive.
  • Go to market was unknowable

Using Custom Controls Got Much Simpler with AI

With the emergence of the AI capabilities (in our case mostly Claude), AI can serve a collaborator to the developer. The two together move faster than either could alone. The ceiling on what a Cognos developer can build in a given engagement just got significantly higher.

The combination of AI Collaborator and a Cognos Developer could work as follows:

  • The AI collaborator brings broad technical range: JavaScript patterns, browser debugging approaches, research across community forums and IBM documentation. It can hold the full context of a complex debugging session and reason across it without losing the thread.
  • The Cognos developer brings the domain knowledge: how the platform behaves at runtime, where data lives in a rendered report, how reports communicate with each other, what the platform expects when one component needs to talk to another. That knowledge cannot be looked up in a manual. It is earned through years of working with the product across real client environments.

What would have taken weeks of solo development -- or significant outside development costs -- required approximately 14 to 16 hours of aggregated working time with AI as a collaborator. And now it is reusable across any report with a simple JSON configuration block.

Example Use Case: Interactive Reporting

We used the custom control framework to build a hover-driven contextual detail control for IBM Cognos Analytics 12.1. When a user hovers over any cell in a list or crosstab report, a popup appears above the report displaying a related detail report and filtered automatically to the data behind that cell. No navigation away from the main report. No manual filtering. The detail appears, the user reads it, the popup dismisses when they move on.

This is the kind of interaction that users of leading visual analytics platforms take for granted. Point at a number, see what is behind it. It is a natural, expected behavior in modern BI.

The result is a first-class interactive element -- indistinguishable to the end user from anything the platform ships natively. The hard work of security, data retrieval, and rendering is already done. The developer inherits all of it.

What This Means for Your Cognos Investment

This can be a better story than migration to the next shiny object. Fortune 500 companies that run Cognos Analytics have something they may not fully appreciate: a mature, governed, enterprise-grade reporting platform that has quietly become extensible in ways that fundamentally change its capabilities and functionality.

The organizations that understand what they have, and find the right partners to extend it, will get significantly more value from a platform they have already paid for and already governed.

Coming in Part Two:

In Part Two we walk through what we actually built: the DOM inspection technique that makes cell-level hover detection reliable across both list and crosstab report types, the promptParameters discovery that solved parameter passing in Cognos 12.1, the architecture decisions that make the control reusable across any report with a simple JSON configuration, and what the debugging process looked like when AI collaboration was part of the workflow.

The code is real. The control is ready for production. The lessons are transferable.

Related Posts

Using watsonx.BI with IBM Cognos Analytics to Streamline Reporting

Watsonx.BI is the new IBM interface to allow users to speak to data. It provides a more intuitive and user-friendly way to understand drivers of...

Read More

Seven Ways to Co-Work with the IBM Cognos Analytics BI Assistant (Part 2)

In today’s fast-paced, data-driven world, organizations that harness the power of AI can boost their efficiency by up to 40%. Imagine having a...

Read More

Cognos Analytics Release 12.0.1 is Here!

WHAT’S NEW IN COGNOS ANALYTICS RELEASE 12.0.1? IBM has introduced Cognos Analytics release 12.0.1and it includes some great new features. To get you...

Read More