Terms you’ll see often on AI Lab

- An AI workflow where the model makes low-level decisions on its own — vetting inputs, choosing which path to take, adjusting the output format — instead of running a single fixed instruction. Why it matters: Jazz's podcast-pitch GPT decided whether each show fit her criteria before drafting a pitch, which is the difference between a one-shot generator and a system you can hand to a virtual assistant.

- A way of describing where a marketer sits on the curve from learning and experimenting to building autonomous systems. Why it matters: the webinar's six stories span the curve, and recognizing your own position helps you choose the next workflow you tackle rather than chasing someone else's setup.

- The practice of feeding an AI enough background — frameworks, voice samples, customer language, past output — that it can make decisions in your voice without being micro-prompted each time. Why it matters: Pam's 142-page document is the canonical example. The quality of the output is bounded by the quality of the context, not the cleverness of the prompt.

- A configurable version of ChatGPT trained on documents and instructions you provide, so it behaves like a specialist on your specific use case. Why it matters: Jazz built one trained on her services, ideal clients, and personal background, then handed it to her virtual assistant to pitch podcasts on her behalf.

- When an AI generates information that sounds plausible but isn't true — often because it lacked the context to know the right answer. Why it matters: Pam called this out as the reason context engineering is non-negotiable; without enough background, the AI fills gaps with confident invention.

- A no-code automation platform (similar to Zapier or n8n) that connects apps and runs multi-step workflows when a trigger fires. Why it matters: Larissa's whole social-content workflow runs in Make, with a Slack message kicking off a sequence that pulls from Notion, calls ChatGPT twice, and writes results back.

- A way of splitting tasks between humans and AI — humans do the maker work (creative, original, generating insight), AI does the mover work (transporting that insight to the right place at the right time). Why it matters: Sean named this as a recurring theme across the six stories — the practitioners kept the judgment work and let AI handle the routing.

- A remote contractor (often hired through platforms like Upwork) who handles repeatable tasks for a small business or solo operator. Why it matters: Jazz's podcast pitch system is built around the VA, not around Jasz — the AI translates Jazz's preferences into outputs the VA can execute.

- Visual workflows in ActiveCampaign that send emails, update fields, and apply tags based on triggers like a form submission, tag application, or date condition. Where it lives: under Automations in the main navigation.

- A feature that shows different blocks of an email to different subscribers based on field values, tags, or AI-generated reasons. Where it lives: inside the email designer, on individual content blocks.

- Saved filters that select a slice of your contact base based on tags, custom fields, or behavior — used to target campaigns and trigger automations. Where it lives: built inside the campaign or automation flow when you choose recipients.

- Contact properties beyond the defaults (name, email) that let you store arbitrary data per subscriber — geography, plan type, contract end date, an AI-generated match reason. Where it lives: under Contacts → Manage Fields.

- Short labels you apply to contacts to mark interest, source, or status — ideal for the "AI-matched this week" pattern Adam used. Where it lives: on each contact's profile, or applied automatically inside an automation.