Segment Zoom webinar attendees into Kit follow-up sequences

Every evening at 6pm on a cron trigger, sweep Zoom for any webinar that finished in the last 24 hours and turn the attendance data into segmented follow-up in Kit (ConvertKit). Build this as an agent workflow, since the segmentation, the recap drafting, and the question triage all need judgement.

Put these settings at the top of the workflow so they are easy to change: the attendance threshold, defaulting to 50 percent of the webinar's actual runtime; the three Kit sequence names, defaulting to "Webinar engaged follow-up", "Webinar recap", and "Webinar replay"; the tag prefix that marks someone as already routed, defaulting to "webinar"; the lookback window, defaulting to 24 hours; and the Slack channel for the summary.

Use Zoom's "List webinars (Accounts)" to find past webinars that ended inside the lookback window. For each one, call "Get webinar detail reports" for the session title and how long it actually ran, "Get webinar participant reports" for every attendee with their email, name, and minutes attended, and "Get webinar absentees" for the registrants who never joined. Also pull "Get webinar Q&A report" and "Get webinar poll reports" so you know what was actually discussed. Note that one attendee can appear as several rows in the participant report if they dropped and rejoined, so sum the durations per email address before doing anything else.

Compare each attendee's total attended minutes against the threshold applied to the webinar's real runtime rather than its scheduled length, since sessions often run long or end early. Attendees at or above the threshold are engaged. Attendees below it are partial. Everyone on the absentees list is a no-show. If somebody somehow appears in both lists, treat them as an attendee.

Before writing anything to Kit, resolve names to IDs with "List Tags" and "List Sequences". Use a per-webinar tag for each segment, shaped like prefix-webinarid-engaged, so rerunning the same session is a no-op while a brand new session still routes normally. Call "List Tag Subscribers" for the three tags belonging to this webinar and build the set of email addresses that were already handled. Skip those people completely: no tag, no sequence, and do not count them as newly routed.

For everyone left, upsert them with "Create Subscriber" using their email address and first name. Kit's V4 API uses email_address rather than email, and Create Subscriber is an upsert that returns 200 when the subscriber already exists and 201 when newly created, so it is safe to call on people already on the list. Then apply the segment tag with "Tag Subscriber by Email" and route them with "Add Subscriber to Sequence by Email": engaged to the engaged sequence, partial to the lighter recap sequence, and no-shows to the replay sequence. Apply the tag before adding to the sequence, so a failure part way through can never leave somebody sitting in a sequence without the tag that would stop a later run from adding them a second time.

Draft a recap email with Kit's "Create Broadcast". Summarize what the session covered, pull the top questions out of the Q&A report and answer them properly, and mention any notable poll results. Leave it as a draft for review: do not set a send time and do not publish it. Include a link to the draft in the Slack summary so it is easy to find.

Post a summary to Slack with "Send a Message" covering the webinar name and runtime, how many people registered, how many attended and the resulting show-up rate, the split across engaged, partial, and no-show, how many people were skipped because they were already routed, and any question that came up more than once in the Q&A. Call out repeated questions and unanswered ones specifically, since those are the ones that deserve a real answer rather than a one-line reply.

Edge cases: if no webinar finished inside the window, do nothing and post nothing. If several webinars finished, process each one separately with its own tags, its own recap draft, and its own Slack message. If a participant record has no email address, which happens when somebody joins without registering, skip that person and report the count in the Slack summary.