The Service Catalogue in 2026: You Cannot Automate What You Have Not Defined

In 2022 we published an article asking whether a Service Catalogue was worth the effort. Our answer then was that it was not really a choice: it is an essential ingredient for any service organisation, IT or otherwise.

Four years on, we still walk into enterprise IT functions, serious budgets, capable people, and find no catalogue at all. No shared definition of the services IT provides. No reference point for what users should expect. Every stakeholder holding a slightly different picture in their head.

What has changed is not the problem. It is the cost of it.

The problems have not gone anywhere

Everything we wrote in 2022 still holds. Without a Service Catalogue you get poor customer experience, because there is no way to set expectations for each service. You get misaligned expectations, because every stakeholder invents their own. You get request prioritisation chaos, because nothing defines what will be fulfilled and when. You get supply chain confusion, because the people contributing to a service cannot see what they are contributing to. And you get broken relationships, between stakeholders and between data: the catalogue is the entry point that connects services to the CMDB, to finances, to suppliers and to the people who consume them.

We described those conversations as nailing jelly to the wall. They still are.

What 2026 adds: you cannot automate what you have not defined

The new pressure is AI. Organisations are making significant investments in AI-enabled operations, intelligent automation, virtual agents and self-service. Nearly every one of those investments quietly assumes a foundation that often does not exist.

An AI agent fulfilling requests needs to know what services exist, who is entitled to them, what good fulfilment looks like and what the business rules are. Automated triage needs services defined well enough to route against. Self-service portals are, in effect, a Service Catalogue with a user interface: if the catalogue underneath is incomplete or wrong, the portal simply industrialises the confusion.

This is the same warning we have given about automation for years: automating a poor process just delivers a poor outcome faster. A missing or unreliable catalogue is the data-quality problem underneath almost every disappointing IT automation initiative we assess.

Why catalogues still fail to get built

The effort involved has not shrunk. Building a catalogue at enterprise scale means multiple stakeholders across countries and cultures, competing perspectives, and a myriad of data sources to collate. The perspective of a senior executive in Singapore is very different from a knowledge worker in the Netherlands, which is different again from a supply chain stakeholder fulfilling locally. That was true in 2022 and it is true now.

What we see most often is not ignorance of the problem but deferral of it. The catalogue is unglamorous foundation work, and foundation work loses budget battles to visible initiatives. The irony is that the visible initiatives, particularly the AI ones, depend on it.

Where to start

Three things make the difference between a catalogue that gets used and one that gathers dust. Agree at the outset which stakeholder perspectives the catalogue serves, before any design work begins. Treat it as a living product subject to continual improvement, not a one-off documentation exercise. And be honest about the current state before you build: in our assessments, Service Catalogue Management is one of the 30 disciplines we score, and it is one of the most common areas where leadership perception and operational reality diverge.

So, is it worth it in 2026? The question has inverted. It is no longer whether you can justify building a Service Catalogue. It is whether you can justify any AI or automation investment without one.