This is the anatomy of a funnel I built for a marketing agency — the one you can read in full as a case study on this site. I'm sharing the architecture because most people building webinar funnels wire the pieces wrong and pay for it in manual work later.

**The pieces.** A landing page captures registrations. Make.com receives each registration and does three things in parallel: creates or updates the contact in ActiveCampaign with correct tags, starts the reminder email sequence, and creates a Google Calendar event the attendee can accept with one click. After the webinar, the attendee list flows back into the CRM so the sales team starts from clean data.

**Where AI actually helps.** People expect AI to write the whole funnel. It shouldn't. The highest-value AI step was drafting personalised reminder and follow-up emails — the sequence now references the webinar topic and the registrant's signup answers instead of a generic email blast. The subject lines get A/B tested by the sequence tool, and drafts that outperform become templates automatically. AI as the last-mile personalisation layer: that's where it earns its keep. AI as the architect of an unattended funnel: mostly noise.

**What I'd do differently.** Two things. First, I'd add the calendar-invite step from day one — I added it later and the no-show rate difference was immediately visible. Second, I'd instrument the landing page with events from the start, because post-facto analytics told us which traffic source produced attendees but not which produced conversions.

**How you know it works.** The client's metric is simple: hours of manual admin per webinar. It went from about six hours to nearly zero, and the funnel has since been cloned for three more webinar series without new development. That reuse is the real test of a good funnel — a second event shouldn't be a second project.

If your webinar (or lead magnet — the architecture is identical) still runs on copy-paste, a free audit will map what yours would look like automated.