See llms.txt for all machine-readable content.
Small business owners, freelancers and solo operators whose Gmail mixes real inbound sales interest with newsletters, vendor pitches and other noise. This workflow uses OpenAI to spot genuine sales leads and labels them in Gmail. It only adds a label; it never sends, deletes or drafts email.
AI/Lead) and optional notes.gpt-4.1-mini by default) decides whether each email is a genuine sales lead (someone who wants to buy from you) and returns strict JSON with is_lead, a short summary and a reason.is_lead is true, the workflow applies your label. If the label does not exist yet, it is created once (with its parent, AI) and then applied.business_name, keep or change openai_model (default gpt-4.1-mini), optionally rename lead_label (the label is created automatically; for a nested label, write its full path, for example AI/Lead), and add plain-English classification_notes if you need them (for example, "Inbound demo requests from the website are always leads").AI/Lead label appears, then flip the workflow to Active.gpt-4.1-mini usually a fraction of a cent per email)lead_label in Settings. Use the full path for nested labels (for example AI/Lead or Sales/Hot). Missing labels are created automatically.max_body_chars to control how much of each email is sent to the model.