Marketing teams have spent years adding workflows, lists, properties, reports, forms, integrations and automations to their martech stacks. The problem is that very little ever gets removed.
Over time, what started as a carefully designed marketing operations environment can turn into hundreds of workflows nobody wants to touch, duplicate reports measuring slightly different things, properties whose purpose nobody remembers, and assets created by employees who left years ago.
A recent HubSpot Admin HUG session featuring Ryan, a Senior Product Manager at HubSpot working on account cleanup and admin automation, highlighted why this is becoming a much bigger issue than simple portal housekeeping.
AI changes the cost of martech clutter
One of the most important points from the session was simple: AI performs better when the context it receives is better.
If an AI assistant is working across a CRM filled with old test lists, duplicate properties, obsolete workflows and inconsistent naming conventions, those assets become part of the environment it has to understand.
That can affect everything from AI generated analysis and segmentation to automation recommendations and reporting.
For Marketing Operations teams, this creates a new reason to care about governance. Cleaning the stack is no longer only about making life easier for administrators. It is increasingly about creating a reliable operational layer for AI.
The biggest problem is often uncertainty
Most experienced MOps professionals will recognize the scenario discussed in the session.
You inherit a portal containing hundreds of workflows. Nobody knows exactly what some of them do. But nobody wants to switch them off either, because something important might break.
So they remain.
Someone cannot find an existing workflow and creates another one. A new report gets built using slightly different fields. A new property is added because nobody knows which existing property should be used.
Gradually, complexity compounds.
The cost appears later through slower campaign execution, inconsistent reporting, difficult onboarding and hours of detective work whenever something breaks.
HubSpot is beginning to automate the cleanup
HubSpot’s Account Cleanup capability is designed to automate parts of this process for Enterprise customers. Admins can define criteria for assets or records that are safe to remove, after which HubSpot can periodically identify and clean up items meeting those rules.
Importantly, the discussion emphasized reversibility and guardrails. Automation becomes valuable only when teams can trust that cleaning the portal will not accidentally remove business critical processes.
HubSpot is also using AI to help organize workflows and segments into folders based on names and descriptions. The longer term direction discussed in the session goes further, including AI assisted naming conventions, duplicate asset detection and greater collaboration with Breeze around portal governance. Some of these capabilities were described as future roadmap areas rather than currently available functionality.
You do not need Enterprise automation to start
Perhaps the more useful lesson for many Marketing Operations teams is that the underlying discipline does not depend on a particular HubSpot edition.
The session suggested several practices teams can implement manually today: regularly reviewing unused workflows and forms, establishing consistent naming conventions, checking duplicate records and low usage properties, reviewing user permissions, and auditing assets created by employees who have left.
One particularly useful idea was the “zero list”: a segment representing a condition that should never occur. If the list suddenly contains records, something in the underlying process may have gone wrong.
That turns cleanup from a periodic project into an early warning system for operational problems.
AI can already help with the audit
The session also discussed connecting HubSpot through its MCP capabilities with tools such as Claude or ChatGPT to help examine the portal and develop a cleanup plan.
The AI can help identify patterns, underused assets and potential cleanup opportunities. However, according to the session, destructive actions remain restricted by safety controls, meaning administrators still retain responsibility for executing many of those changes.
That distinction matters.
The near term opportunity is not to hand your martech stack completely over to an AI agent. It is to use AI as an intelligent assistant that helps administrators understand increasingly complex systems faster.
The RightWave perspective: Treat martech hygiene as infrastructure
At RightWave, we believe Marketing Operations teams should start thinking about martech hygiene the same way technology teams think about infrastructure maintenance.
The objective is not simply to have a “clean HubSpot.”
The objective is to maintain a marketing system where teams can confidently understand what data means, which automations are active, how campaigns are connected, who owns critical processes and which assets can safely disappear.
That foundation becomes even more important as AI starts participating in campaign creation, segmentation, reporting, lead management and other operational workflows.
Before asking AI to automate more of your marketing operations, it may therefore be worth asking a simpler question:
How much of your existing martech environment would you actually want an AI system to learn from?
If the answer makes you uncomfortable, the next AI initiative may need to begin with cleanup.
Reference – https://youtu.be/pozBsQ9az3M?si=Ds93-HAAkM_VgJgi
