Weekly feature adoption report from Amplitude to Notion

By General Input

Every Monday morning, an AI product analyst reads last week's Amplitude usage, writes a narrative per launched feature, and files it in your Notion product review database.

Integrations

  • Amplitude
  • Notion

Type

Agentic Task

Categories

  • Product

Every Monday at 9am (operator-configurable timezone), generate a written feature-adoption report for our recently launched features and drop it into a Notion database as a new page. Trigger: cron. Integrations: Amplitude and Notion.

Let the operator configure: (1) a list of feature event names to track (or a single event category), (2) the ordered funnel steps that represent activation for those features (2 to 4 events), (3) the target Notion database ID where the report page should be created, and (4) the Amplitude user property to segment by, such as plan tier or platform. Provide sensible defaults but treat these as first-class inputs.

For date ranges, compute three windows in Amplitude Dashboard format (YYYYMMDD with no separators): last week (the seven days ending yesterday), the week before that, and the week before that. The two prior weeks are the trajectory baseline.

For EACH configured feature event, call Amplitude Get Event Segmentation Data for last week, broken down by the configured user property (plan tier or platform). Then call Amplitude Get Event Segmentation Data for each of the two prior weeks so you can compare totals and detect trajectory (growing, flat, declining). Also call Amplitude Get Funnel Analysis Data once per feature using the operator-supplied funnel step sequence over last week to compute activation conversion, and again for the prior week to see if the funnel is improving or getting worse.

Synthesize a short adoption narrative for each feature (roughly 4 to 6 sentences): who is using it (top segments by the configured property), who isn't (segments with near-zero usage or a sharp drop-off in the funnel), whether the activation funnel is converting (cite the last-week conversion rate and the delta vs the prior week), the two-week trajectory, and ONE concrete follow-up action a PM could take this week (for example: reach out to the underperforming segment, tighten the empty state on the drop-off step, or promote the feature to power users who haven't tried it).

Then call Notion Create a Page under the operator-configured Product Reviews database. Title the page with the ISO week (for example 'Feature adoption — 2026-W28'). In the page body, include one section per feature with: the narrative, a small table of last-week totals broken down by segment, the funnel conversion rate with the week-over-week delta, and the follow-up action. Add a short executive summary at the top listing which features are trending up, which are flat, and which need attention.

Notes for the agent: Amplitude Dashboard date parameters are strict YYYYMMDD with no dashes. If a feature event has zero volume across all three weeks, still include a one-line entry in the report noting no signal was captured — do not silently omit it. If the funnel returns fewer than the expected step counts, flag the gap explicitly in that feature's narrative rather than fabricating a conversion number.

Related prompts

Explore more prompts
Screen chartered vessels for AIS gaps and sanctions red flagsCompare candidate store locations with drive time trade areasBrowse every competitor ad and build a Notion swipe fileAudience persona studio for brand and product marketersArtwork desk for the Notion posts still missing an imageShare of voice dashboard for your brand and competitorsInfluencer campaign roster board with AI creator briefsSmart home entity health cockpit for Home AssistantOpen model approval desk for your production AI stackBid board that watches every procurement portal for you