AAgency MCP
Dashboard Permissions Pending Actions Documentation Support
Privacy Terms
AI tools can make mistakes. Always grant the minimum scope to reduce risk. Your working data isn't used to train our models.

Roadmap

This page describes the broad direction of Agency MCP over the next several weeks. Dates are targets, not guarantees. We ship a marketplace "What's new" update for every release.


Near-term focus: safer write operations

The next major investment is reducing the blast radius of AI-initiated changes to your GHL account. Today, AI agents that use Agency MCP can make changes directly. The next release introduces an opt-in mode where write operations are reviewed before they execute — your AI surfaces what is about to change and waits for your approval (or a configured automated policy) before committing.

This lets you grant your AI wider autonomy without worrying about unintended mass changes. The first workflows shipping under this model focus on contact and tag operations.


Continuous tool guidance

Your AI's effectiveness depends on knowing which Agency MCP tool to use, when, and with what parameters. We are investing in a process to improve tool guidance continuously — adding examples, noting quirks, suggesting better alternatives. These improvements reach your AI client automatically; you do not have to reconfigure anything.


Operational awareness

Several upcoming features sharpen your view into what is happening in your account:


Beyond the near term

Longer-horizon work depends on what testers ask for. Likely candidates:

Per-bearer scope permissions are already shipped — see Scope permissions.

If a specific capability matters to you, tell us on the Support page. Tester feedback shapes what we build next.


Subscribe to our Discord for release announcements. The link is on the Support page.