A morning news briefing worth opening

In the version I shared on LinkedIn, the report ran at 7 a.m. and scanned the previous 48 hours of Google News. It covered roughly 40 of Ordergroove’s top customers, direct company mentions, and relevant industry news. On Fridays, it searched a broader list of 100 accounts.

Each article received a score for timeliness, source quality, and other relevance checks. Strong matches stayed in the digest. Borderline matches went into a review section, and junk dropped out. The workflow read the selected articles and wrote a one-sentence recap with a separate explanation of why the story mattered for Ordergroove.

I received the digest in a Slack DM with a link to the full dashboard. It caught a merchant expanding subscriptions into another brand and a couple of quiet mentions of Ordergroove’s role in subscription growth. The expansion gave us a reason to start a broader partnership conversation.

The downloadable prompt walks you through a guided interview, then creates a research prompt from your answers. It asks about your role, topics, sources, freshness window, and the kinds of stories you want. It also asks what to leave out. That last part deserves some thought.

You review the proposed topic groups and summary before it generates the final prompt. A marketer watching customer news should get a different briefing from an investor tracking funding rounds.

1. Download the prompt and paste it into your AI tool. Answer the intake questions with the actual work you need the briefing to support.
2. Review the proposed topics and story counts. Start with a list you will read, even on a busy morning.
3. Run the finished research prompt in a tool with web access. Open the source links and check dates, relevance, and whether each summary matches the article.
4. Refine the prompt after a few runs. Set up scheduling and delivery separately once you like what comes back.

This is the prompt builder I shared, including its questions and final research-prompt template. It asks for primary sources within your chosen time window. It does not include the code for my scoring system, dashboard, or Slack delivery.

The schedule you choose becomes part of the generated prompt. You still need to configure a scheduler to make it run automatically. Keep checking the output: an instruction to verify a source does not prove that the tool did it.