AI enrichment rows may time out due to rate limits on your API tier, especially for heavier prompts or large-scale runs. Claygent web research requires a Tier 2 OpenAI API key (≥450,000 TPM), while ChatGPT text generation needs 30,000 TPM. If your API key is on a lower tier, actions may run slowly or fail. To resolve this, upgrade your API tier and break large requests into smaller batches. Also, ensure your prompts and column configurations are specific and standardized for consistent results. [1]