Turn webinar feedback into three concrete improvements

**Objective:** Summarize what the audience said about the webinar `{webinar-name}` in the session on `{session-date}` (default: the most recent session that has already happened), from its polls and its post-session survey.

**Step 1: Identify the webinar and session**

Call `goto-webinar-get-webinar` with webinar-name.contains set to `{webinar-name}`. If more than one distinct webinar key matches, list the matches (name, date, webinar key) and stop. Call `goto-webinar-get-session` with that webinar key and pick the session dated `{session-date}`, or if that is empty the latest session on or before today.

**Step 2: Aggregated results**

* Call `goto-webinar-get-sessionsurvey` with the session key.
* Call `goto-webinar-get-sessionpoll` with the session key.

For each question, collect every response with its times-answer-chosen and percentage-chosen, and the question's total-responses.

**Step 3: Open-text answers**

Call `goto-webinar-get-attendeesurvey` with the session key and max-records 1000. Keep the answers to questions whose question type is free text (short or long answer, not single or multiple choice). Group them into themes.

**Step 4: Report**

* Response rate: the largest total-responses on any survey question against the session's registrants-attended.
* Rating and choice questions: a table of Question, Leading response, Percentage, Total responses. Call out any question where the leading response is negative.
* Polls: the same table, kept separate from the survey.
* Open-text themes, largest first, each with its count and up to two short verbatim quotes.
* Three actions for the next session, each tied to a specific result above.

**Rules & Guidelines:**

* This prompt is read-only. Never call `goto-webinar-add-registrant` or `goto-webinar-delete-registrant`.
* Do not put names or email addresses next to quotes. Feedback is reported anonymously.
* With fewer than 10 responses to a question, say the sample is small rather than drawing a conclusion from it.

**Validation:**

* If the session had no survey and no polls, say so and stop. That is not negative feedback.
* If no session qualifies, say so and list the webinar's session dates.