Weekly customer review digest in Slack and Notion

Build me an agent workflow that runs every Monday at 7am ET on a cron trigger and produces a weekly Voice of Customer reputation digest from public product review sites.

Inputs I will configure: a list of product review page URLs (a mix of G2, Trustpilot, Capterra, Apple App Store, and Google Play listings), the Notion database ID for my Voice of Customer database, and the Slack channel ID where the digest should land.

For each URL in the list, the agent should use Browserbase to scrape the reviews. Specifically: call Browserbase Create Session to provision a headless browser session and get its connect URL, then drive that session with Stagehand or Playwright over the returned connect URL to load the review page, dismiss any cookie or login modals, scroll until reviews older than seven days come into view, and extract every review posted in the last seven days. For each review capture the star rating, the date, the reviewer role or company segment when visible (for example job title, company size, industry), and the full review text. When the session is done, call Browserbase Release Session so the session does not sit idle until timeout.

If a given source has zero new reviews in the last seven days, skip it. Do not pad the digest with empty sections.

Once all reviews are collected across sources, use the agent's judgement to cluster them into themes (examples: onboarding friction, pricing complaints, new feature praise, performance issues, support quality). For each theme, pick the two most actionable quotes, with attribution to the source site and reviewer role when available. Also flag any single review that looks like a churn risk or a public escalation, for example a one star review from an enterprise reviewer, a threat to switch tools, or a public callout that could spread.

Write the results to Notion using Create a Page in my Voice of Customer database. The page title should be the week range (for example "Voice of Customer, May 26 to Jun 1, 2026"). Page properties and body should include the overall sentiment for the week, total review count by source, the theme breakdown with quote highlights, and a dedicated section for any churn risk or escalation flags. Keep the structure consistent week over week so entries are comparable.

Then post a short Slack digest using Send a Message to the configured channel. The Slack message should lead with this week's overall sentiment, list the top three themes with one line each, call out any churn risk flags, and link to the new Notion page for full detail. Keep the Slack message scannable, not a wall of text.

Integrations to use: Browserbase (Create Session, Release Session) for scraping the review pages, Notion (Create a Page) for the weekly entry, and Slack (Send a Message) for the digest.