Weekly Brevo campaign autopsy with recommendations in Slack

Every Monday at 9am, autopsy last week's Brevo email campaigns and post a recommendation-driven recap to a marketing Slack channel so the team knows what to do differently this week. This is an agent workflow, not code, because the value is the qualitative pattern-finding and recommendation writing, not the data pull.

Trigger: cron, weekly on Mondays at 9am in the org's local time zone.

Step 1. Pull last week's campaigns. Call Brevo's List Email Campaigns with status=sent and filter to campaigns whose sentDate falls in the last 7 days. For each campaign, capture the campaign name, subject line, sent date and time, recipients, and the per-campaign open rate, click rate, bounce rate, and unsubscribe rate that come back on the campaign object. Then make a second List Email Campaigns call covering the prior 7-day window (days 8 through 14 back) so you have a comparison baseline.

Step 2. Rank. Compute the prior week's median open rate and median click rate from the baseline pull. Tag last week's campaigns as top performers when they beat both medians and as underperformers when they miss both. Sort top performers by click rate descending and underperformers by click rate ascending.

Step 3. Find patterns. Look at the winners and call out qualitative patterns that correlate with the wins. Examples: subject-line themes or formats (questions, numbers, personalization, length), topics, audience segments, and send-time windows (day of week and hour of day). Be concrete and quantitative where you can, for example "Tuesday 10am beat Thursday 3pm by 18 points open rate" or "subject lines mentioning [topic] averaged a 32% open rate vs 19% for the rest".

Step 4. Write 2 or 3 concrete recommendations for the upcoming week, grounded in the patterns. Each recommendation should be a single sentence telling marketing what to do differently, for example "lean into [topic] subject lines" or "shift the midweek send from Thursday afternoon to Tuesday morning". Avoid vague advice like "keep testing".

Step 5. Post to Slack. Use Slack Bot's Send a Message to post to the marketing channel the user picks at setup. Format with Slack mrkdwn (single asterisks for bold, not double). Structure the message as: a one-line headline summarizing the week (for example "5 campaigns sent, opens up 4 points vs last week"); a *Top performers* section listing each winner with its open rate, click rate, and subject line; an *Underperformers* section with the same fields; a *What we noticed* section with the patterns; and a *Recommendations for this week* section with the 2 or 3 action items as a bulleted list.

Edge cases. If zero campaigns sent in the last 7 days, post a short "no sends last week" note and skip the rankings. If only one campaign sent, skip the ranking step and just summarize that campaign plus one forward-looking recommendation. If the prior week's baseline is also empty, compare against the campaign's own historical averages from the campaign object instead of a weekly median.

Integrations: Brevo (API key) and Slack Bot (OAuth). The user will pick the Slack channel during setup.