# Nobody Wants to Browse Your Catalog

_Ran Ribenzaft · 2026-08-04_

**Categories:** ITSM, CIO, IT Service Management

Why catalog items and legacy portals are absurd in 2026 - for the employees who use them and the IT teams who maintain them

It's 9:40 on a Tuesday morning. A designer needs access to Figma for a contractor starting Thursday. She opens the IT portal, types "Figma" into the search bar, and gets three results: "Software Access Request (SaaS)," "New Vendor Application Review," and "License Provisioning - Design Tools (Legacy)." She picks the second one, because it mentions vendors and this is for a contractor. Wrong. The ticket bounces back two days later with a note: please resubmit using the correct catalog item.

Ten minutes earlier, that same designer asked an AI assistant to restructure a project brief, and it just… did it. No form. No dropdown. No guessing which of three near-identical items maps to her intent.

That gap - between how people work all day and how they're forced to ask their own company for help - is the most visible absurdity in enterprise IT right now. And the catalog item sits at the center of it.

**The form is doing the computer's job**

Let's be honest about what a catalog item actually is. It's a workaround from the mid-2000s. Software back then couldn't understand a request written in plain language, so we did the next best thing: we made humans pre-structure their requests into something a database could swallow. Category. Subcategory. Assignment group. Urgency. Business justification. Twelve fields, two attachments, and a dropdown with 40 options that hasn't been updated since the last reorg.

Look at those fields one by one and ask a simple question: who already knows the answer?

Who is making the request? _Your identity provider knows. _What device are they on? _Your MDM knows. _Who is their manager, what department, what location? _Your HRIS knows. _What's the right category and priority? _IT decides that anyway - usually by overriding whatever the employee guessed._

Nearly every field on a catalog form is a question the organization can already answer without asking. The form isn't collecting information. It's outsourcing data entry to the least qualified person in the process: the employee who just wants their problem solved. And because employees know the fields are a formality, they fill them accordingly - which is why so many tickets arrive with the wrong category, "high" urgency, and "see title" in the description.

**On the other side of the portal, it's worse**

The employee experience is the visible half of the absurdity. The invisible half is what catalogs cost the teams that maintain them.

Every catalog item is a small software project. Someone designs the form. Someone builds the field validation, wires the approval chain, maps the assignment groups, tests the flow, and documents it. Then the org changes - a team splits, a tool gets replaced, an approver leaves - and every one of those items quietly drifts out of date. Onboarding a new department onto the service desk takes weeks or months, not because the work is hard, but because the work is forms.

Talk to any IT lead running a mature ITSM instance and you'll hear the same confession: nobody prunes the catalog. Items accumulate like sediment. The portal offers three ways to ask for the same thing and no way to ask for the newest thing, and the team that should be improving service is instead chasing fields, approvals, and stale dropdowns. Every change is a project; the backlog of portal fixes competes with the backlog of actual work.

We built a system to reduce IT's manual effort, and it now consumes IT's manual effort. That's the tell that a technology has outlived its constraints.

**2026 broke both assumptions**

The catalog rested on two assumptions. One: software can't understand what a person is asking for, so a human must structure the request. Two: software can't gather context on its own, so a human must supply it. Both were true in 2006. Neither is true in 2026.

Language models understand "my contractor starts Thursday and needs Figma" better than any form ever captured it - including the parts no form would think to ask, like the deadline. And context no longer needs to be typed in, because it's sitting in the systems you already run: the IdP knows who's asking, the HRIS knows their role and manager, the MDM knows their device, the knowledge base knows the policy. An agent that can read those systems doesn't need the employee to fill twelve fields. It needs one sentence.

Meanwhile, your employees' expectations have been reset by the tools they use all day. They live in Claude and ChatGPT, describe what they need in a sentence, and get resolution in seconds. Then they hit a 20-year-old portal that asks them to classify their own ticket. The portal didn't get worse. Everything around it got better, and now the contrast is impossible to ignore.

**What replaces the catalog isn't a better catalog**

The instinct in our industry is to modernize the form - fewer fields, nicer UI, maybe an AI that helps you pick the right catalog item. That's polishing the workaround. If the software can understand the request and gather the context, the right number of catalog items to show an employee is zero.

Here's what that looks like in practice. The request starts as a sentence, in Slack or Teams - where people already work. The agent enriches it with identity, device, and access context automatically. A properly tagged, prioritized, routed ticket gets created behind the scenes, synced two ways with the ITSM you already run - the system of record stays intact; the form disappears. Common requests are resolved end to end by pre-built skills that work on day one, and the processes unique to your org get described in plain language and become working agents in minutes - not built as form-and-workflow projects over weeks.

This isn't speculative. At Harmony we're seeing 65% of employee requests resolved with no human touch within the first 48 hours of deployment, climbing past 90% as agents learn the organization - with employees rating the experience 4.85 out of 5. It turns out people don't hate asking IT for help. They hate the portal standing between them and the help.

**The catalog was never the product**

None of this is a knock on the people who built and maintained these portals. The service catalog was a genuinely good answer to the constraints of its era - it brought order to chaos when the alternative was an inbox. But the form was never the point. Resolution was the point. The catalog item was scaffolding around software that couldn't understand people, and that scaffolding has been load-bearing for so long we've mistaken it for the building.

In 2026, an employee should be able to say what they need, in their own words, in the tool they're already in - and have it resolved. Everything else is legacy.

**The best catalog item is the one nobody ever has to find.**