Airbyte has introduced customizable workspaces that enable administrators to create isolated instances with granular access controls for users, data connectors, and AI agents. Each workspace operates independently with its own governance rules, user permissions, and billing tracking, allowing organisations to enforce use case-specific data access without compromising security across the platform. The feature is available on Airbyte's Team plan at $299 monthly, with additional agent operations priced at $0.005 each. Alongside this, Airbyte has expanded its HubSpot Agent Connectors to allow AI agents to create and update contacts, companies, deals, and tickets—with administrators retaining control over what data agents can access and modify.
For CX teams already managing multiple departments or customer segments through a single platform, this workspace architecture addresses a critical operational challenge: how to grant different teams access to different data without creating security vulnerabilities or governance nightmares. Support leaders running Zendesk or Salesforce integrations will recognise the appeal—the ability to isolate a sales team's deal data from a support team's ticket data, whilst allowing both to leverage AI agents within their respective boundaries, reduces the risk of accidental data exposure and simplifies compliance audits. The HubSpot connector expansion is particularly relevant for teams using HubSpot as a CRM backbone, as it moves AI agents from read-only observation into active data modification, raising the question of whether your current approval workflows and audit trails can handle autonomous agent actions at scale.
The pricing model and governance controls suggest Airbyte is positioning itself for enterprise adoption, but the real test for CX professionals will be integration friction. Does this workspace isolation play cleanly with your existing Zendesk or Salesforce connectors, or does it introduce another layer of configuration overhead? For teams already stretched thin managing multiple platforms, the promise of workspace-level governance only delivers value if it doesn't require rebuilding your entire data pipeline architecture.
Companies can create multiple workspaces with access controls for users and AI agents, allowing different data governance depending on the instance.