We are experiencing an issue with HubSpot’s system intelligence updates overwriting previously revised context in the new AI Context page. In our case, the system reintroduced a service line that we no longer offer, even though that information had already been removed from earlier AI settings.
Based on guidance from support, system intelligence pulls from multiple sources when generating updates. As a result, even when a company has clearly communicated a divestiture through channels such as press releases and website updates, the system may still restore outdated information if legacy references appear more frequently across its source set. This creates a risk that older or no-longer-valid positioning will continue to resurface despite intentional updates made by the customer.
I would like to submit an idea for an exclusion or override capability within AI Context. This functionality would allow customers to identify specific services, terms, phrases, or positioning that should not be added, restored, or suggested by system intelligence. A control like this would help customers better train and govern the system’s updates while protecting the accuracy of their approved context.
Without this type of safeguard, system-generated updates may continue to reintroduce outdated services or messaging, which could directly affect downstream tools such as Prospecting Agent and marketing-related agents.