AI Usage Policy

Oddyssey runs the work of a marketing department as agents: research, strategy, creative production, funnel building, media buying and attribution analysis. They draft hypotheses, briefs, copy, creative, pages and campaign structures, and they read results back against the variable you declared.

• Spend money without an authorized budget ceiling and an approval from an account that holds spend authority.
• Publish a page, launch a campaign or send a message to your customers without the approval step you configured.
• Change your payment gateway, payout settings or connected account credentials.
• Invent a claim about your product. Claims must come from what you provided.

Approval is a product feature, not a promise in a policy. You configure who can approve what, and which actions require an approval before execution. If you widen those permissions, you widen what can happen without a human reading it first. That choice, and its consequences, are yours.

Generating output means sending your prompts, briefs, brand material and the context you attach to third-party model providers through our AI gateway. We do not permit those providers to train their models on your content. We do not use your content to train models of our own.

Do not paste customer payment details, government identifiers or other sensitive personal data into a prompt. Agents do not need it and it does not belong in a generation context.

Model output can be confidently wrong. It can invent a statistic, misread a regulation, restate a claim you never approved or produce copy that a platform will reject. Oddyssey adds structure around that risk, including declared variables, claim checks against what you supplied and decision rules set before launch, but structure reduces the risk rather than removing it.

Review anything before it goes live. Legal, medical, financial and safety-related copy needs a qualified human, not an agent.

As between you and us, you own the output generated in your workspace and you are responsible for what you publish, send and spend. Generated material may resemble output produced for someone else from a similar prompt, and copyright in AI-generated content is unsettled in many countries, so do not assume exclusivity over a generated asset.

Some platforms require you to disclose AI-generated or AI-altered creative, particularly for people and voices. Meeting that obligation is yours. Features that generate synthetic presenters or voices are not released; when they are, they will ship with disclosure controls and their own terms.

We log agent actions, their inputs and their approvals so you can see why something happened. These logs are part of your workspace data and are covered by the Privacy Policy and the Data Processing Addendum.

Model providers and capabilities change. We will keep this page current and note material changes to what agents are allowed to do. Questions go to support@tryoddyssey.com.