1. Where QueueTempo uses AI
QueueTempo can use an AI model to create or regenerate platform-specific text Variants from a Content recipe, Brand profile, selected Knowledge sources, recent content, destination constraints, and your instructions.
AI assistance is optional for manual writing and editing. QueueTempo V1 does not generate video, clone voices, create avatars, make automated direct messages, or claim to predict reach, engagement, revenue, or follower growth.
2. What the model receives
A generation request may include the minimum customer-selected context needed for the requested Variant, such as:
- approved Brand profile fields and factual claims;
- the selected Content recipe and platform constraints;
- approved Knowledge-source excerpts with provenance;
- recent content needed for duplicate avoidance;
- the requested tone, locale, link, call to action, and formatting rules.
Customer Knowledge sources are treated as untrusted input. Instructions embedded in imported material do not override QueueTempo’s system, safety, tenant, or product rules.
3. Output limitations
AI-generated text can be wrong, stale, incomplete, biased, repetitive, or unsuitable. It may produce language similar to output created for others. A structured response and runtime validation improve consistency but do not prove truth, rights, legality, or platform acceptance.
You are responsible for confirming factual claims, permissions, disclosures, professional suitability, platform rules, and the final publication. Keep Review required enabled until you trust the Organization’s context, Content recipes, Connections, and operational controls.
4. Sensitive content
Health, financial, legal, political, employment, housing, education, and regulated-goods content may require verified sources, disclosures, stricter moderation, Review required, or disabled Auto-publish. QueueTempo does not claim that an AI model or automated classifier makes content compliant with law or professional obligations.
5. Validation and moderation
A draft can be checked for:
- platform length and format;
- required and prohibited phrases or topics;
- moderation and high-risk content;
- unsupported factual claims;
- safe links and eligible media;
- near-duplication of recent content.
A failed check should block or skip publication visibly. QueueTempo does not silently weaken a rule to force a post through.
6. Provenance and history
QueueTempo records enough provenance to explain which model configuration, prompt version, Brand-profile version, Content-recipe version, and source versions produced a draft. This is operational provenance, not hidden model reasoning.
QueueTempo does not expose, store in customer logs, or claim access to private chain-of-thought. It stores safe structured outputs and the product events needed for validation, support, billing, and audit.
7. AI credits
One AI credit produces one usable platform-specific text Variant. Creating one X draft uses one credit; creating one Instagram caption uses one. Adapting a logical Post to both uses two. Regeneration uses another credit when it produces another usable Variant.
QueueTempo reserves credits before generation, consumes them when a usable Variant is stored, and releases them when a provider, network, internal, or policy-blocked failure produces no usable Variant. Manual writing and editing use zero AI credits.
8. AI providers and customer data
QueueTempo’s AI subprocessor is OpenAI, called from the server. The Privacy Policy and its subprocessor list describe the provider, data terms, region, retention, and any optional controls in use.
QueueTempo does not use customer prompts, private brand facts, generated drafts, unpublished posts, uploaded media, or private analytics as QueueTempo marketing inventory. QueueTempo does not use that content to train a QueueTempo-owned general model unless the customer separately and explicitly opts in to a future, clearly described program.
9. Auto-publish is a separate decision
Generating a draft does not activate a Tempo or authorize publication. OAuth also does not authorize QueueTempo Auto-publish. An authorized member must expressly consent after seeing the destination, cadence, sample, and approval policy.
10. Questions or corrections
If a draft contains an unsupported claim or a source is outdated, correct the Brand profile or Knowledge source and regenerate or edit the draft. Report repeated safety or privacy concerns through QueueTempo Support without including tokens or unrelated private content.