AI Agent Actions
Chuhching lets you deploy AI agents that act inside your workspace. This section governs those actions. Capitalised terms not defined here have the meaning given in the Terms of Service.
A. Authorization to act on your behalf
You authorise the Service and the AI agents you deploy to take actions on your behalf, within the permissions you configure. Those actions may include, without limitation, drafting and sending messages, updating records, publishing content, scheduling, and initiating payments or transfers up to the budgets you set. An action taken by an agent within your configured permissions is treated as an action you authorised.
B. Approvals and auto-approval are your configuration
The Service can require your explicit approval before an agent performs a sensitive action. Where you approve an action from an approval queue, that approval is your decision. Where you configure a category of actions to run automatically without individual approval, that configuration — and every action taken under it — is your choice and your responsibility. You are responsible for reviewing and maintaining your approval settings and budgets.
C. What we provide, and what remains yours
Chuhching provides the guardrails: configurable spend budgets, approval queues for sensitive actions, per-agent and per-workspace permission scoping, an audit trail of agent activity, and a kill switch. You remain responsible for:
- the instructions and configuration you give your agents;
- the third-party accounts and integrations you connect, and your authority to use them;
- the content, accuracy, and legality of anything an agent sends, publishes, or transmits on your behalf, including compliance with applicable messaging and marketing laws such as the CAN-SPAM Act and the Telephone Consumer Protection Act (TCPA); and
- obtaining and maintaining any consents required for the recipients you contact.
This section supplements, and does not limit, your obligations under the Acceptable Use Policy and outreach compliance guidance.
D. Allocation of liability
Chuhching is responsible for defects in the platform itself — for example, a guardrail that fails to operate as documented, or an agent that exceeds the permissions you configured because of a platform fault. Subject to the limitations of liability in the Terms of Service, Chuhching is not liable for actions an agent takes at your direction and within the permissions you configured, including actions taken under an auto-approval setting you enabled. The overall limitations and exclusions of liability in the Terms of Service apply to this section in full.
E. Kill switch and incident cooperation
Chuhching maintains a kill switch that can halt agent activity, and we may use it to stop activity we reasonably believe is causing harm, abuse, or a security incident. You may also stop your own agents' activity at any time using the controls in the Service. In the event of a security incident affecting agent actions in your workspace, we will cooperate with you in good faith to investigate and remediate, consistent with our commitments in the Data Processing Agreement.
Questions
Questions about this section? Email legal@chuhching.com.