What “AI” means here
When we refer to AI, we mean software systems (including large language models and other machine learning models) that generate or transform text, structured data, recommendations, or creative variants based on inputs you provide and data available through integrations you authorize.
Human oversight
AI outputs are assistive. You remain responsible for decisions that affect your brand, legal compliance, budgets, and published advertising. You should review outputs for accuracy, claims substantiation, and suitability before use in production campaigns.
Model providers and subprocessors
We may route prompts and context to third-party model providers and hosting vendors to generate responses. Those providers process inputs under their infrastructure agreements and security programs. We select providers and configurations with privacy and reliability in mind, but we do not control third-party systems.
Accuracy and limitations
Models can hallucinate, omit important constraints, reflect training biases, or misunderstand niche terminology. Performance may vary by language, region, and prompt quality. Do not rely on AI outputs as legal, financial, medical, or professional advice.
Training on your data
Unless we explicitly notify you otherwise for a specific feature, we configure commercial AI services to use your customer content for the purpose of providing outputs to you — not for public model training. Provider terms may still apply; review provider documentation for details.
Logging and improvement
We may retain prompts, outputs, and related metadata as described in our Privacy policy for security, debugging, quality improvement, and legal compliance. Where feasible, we minimize sensitive content and apply access controls.
Safety and abuse
We may block, filter, or rate-limit requests that appear malicious, violate policy, or risk harm to users or systems. Automated moderation is imperfect; appeals can be handled through support.