Legal

DUMO AI and Agents Terms.

Last updated: 20 November 2026. Applies to dumocloud.com and every DUMO app.

The short version.DUMO AI proposes and you confirm. Agents act only within the permissions you give them and ask before anything important. AI can make mistakes, so check what matters. Your content stays yours and isn’t used to train models for anyone else.

These terms add to the Terms of Service for Ask DUMO, the DUMO agents and every AI feature in DUMO (together, DUMO AI).

1. What DUMO AI does

1.1 Ask DUMO answers questions, drafts content, summarises and, when you ask, does work in your workspace (for example creating a booking, drafting a campaign or filling in a record).

1.2 Agents are AI assistants you hire into your workspace for a role, such as answering customers in the One Inbox, following up leads or keeping your schedule full. Each agent works from the instructions, knowledge and permissions you set.

2. You stay in control

2.1 DUMO AI acts only within the permissions of the person or role it acts for. It can’t see or do anything that person couldn’t.

2.2 DUMO AI asks for your confirmation before it sends a message on your behalf for the first time in a conversation type you haven’t approved, takes or refunds a payment, deletes anything, changes settings or permissions, or uses more than 3 credits on a single action. You can require approval for more actions in agent settings.

2.3 You can pause, edit or remove an agent at any time. Every action an agent takes is logged with what it did and why, and you can review and undo actions where the underlying feature allows.

2.4 Agents hand a conversation over to a person when they’re unsure, when a customer asks for a human, or when your rules say so.

3. Accuracy and your responsibility

3.1 AI output can be incomplete, out of date or wrong, and can sound confident when it’s wrong. Review it before you rely on it, especially for legal, financial, tax, medical or safety matters, which DUMO AI doesn’t give professional advice on.

3.2 You’re responsible for what you or your agents send or publish, for the instructions and knowledge you give them, and for using DUMO AI in line with the law and the Acceptable Use Policy.

3.3 Telling people they’re talking to AI. Where the law requires it, or where a customer might reasonably think they’re talking to a person, you must make clear that an AI agent is replying. DUMO labels agent replies by default; don’t remove the label where it’s required.

3.4 Don’t use DUMO AI to make decisions about people (such as credit, hiring, housing or access to essential services) that have legal or similarly significant effects without meaningful human review.

4. Your content and AI

4.1 Inputs and outputs are yours. As between you and us, you own what you put in (prompts, files, instructions) and, to the extent the law allows, what DUMO AI produces for you. Similar requests from others may produce similar output, which isn’t exclusive to you.

4.2 No training on your content. We don’t use your inputs, outputs or workspace content to train AI models offered to other customers. We may use de-identified, aggregated information about how features are used (not the content itself) to improve DUMO.

4.3 AI providers. DUMO AI runs on models from third-party providers. They process your content only to return a result to us, are contractually barred from training on it, and are listed as subprocessors in the Data Processing Addendum.

5. Credits

Ask DUMO and agent work use credits from your wallet at the rates on the pricing page. DUMO shows the cost before you confirm larger jobs and asks first before any single action uses more than 3 credits. You can set spending limits and alerts in Billing.

6. Availability and changes

We may change the models behind DUMO AI to improve quality, safety or cost. Features may be limited during high demand. Agents marked “Coming” or “Beta” are previews.

7. Reporting problems

If DUMO AI produces something harmful, offensive or plainly wrong, use the report button on the reply or email [email protected]. We review reports to improve safety.

Questions about this page? Email [email protected].