💰 Business recipe
Daily Sales Pipeline Report AI Routine for Mac
A daily sales pipeline report from your CRM: deals that moved, deals that went quiet, and what to chase next.
The exact prompt
Paste this into a new routine.
Open my CRM in the browser, log in with my saved credentials, and pull my current pipeline. Summarize deals that moved stage since yesterday, deals that have gone quiet with no activity in over a week, total forecast for this month, and the three follow-ups most worth making today. Rank by deal value and flag anything at risk of slipping.
Set the schedule. Keep the default weekday 8:00 AM so the pipeline read is ready before your first call. cron: 0 8 * * 1-5
Set it up
Step 02Set up Browser Automation
One-click setup spawns a Playwright MCP server using system Chrome, with your login saved locally, so Routines can read CRMs that have no easy API.
Step 03Pick the CRM Pipeline Summary recipe
Choose Daily CRM Pipeline Summary from the library, or paste the example prompt into a new routine.
Sample output
What a run looks like.
Saved to your local output history · stays on your Mac
Pipeline · 8:00 AM
Done
Advanced (2): Acme to Proposal, Globex to Negotiation.
Gone quiet (3): no touch in 8+ days, follow-ups suggested.
At risk: the Q2 renewal is slipping, chase today.
- Who it's for
- Founders and sales leads who want a daily read on the pipeline without logging into the CRM and clicking through every stage.
- What lands in your morning
- A weekday summary of your sales pipeline: which deals advanced, which went quiet, what is forecast to close, and the specific follow-ups worth making today.
FAQ
Questions about this recipe.
Which CRM does this work with?
Most web-based CRMs. Because it reads your CRM through the Playwright browser automation connector rather than a dedicated API, it can pull from CRMs that have no convenient API. If your CRM exposes a Model Context Protocol server, you can connect that instead for a cleaner integration.
How does it log in to my CRM?
The browser automation connector logs in with credentials saved locally on your Mac, in a Playwright session that stays on your machine and never reaches a BinarCode server. The pipeline data it pulls out goes to the AI to be summarized, through Routines AI, which stores none of it, or straight to your provider on your own key.
Why does this recipe use a higher intelligence tier?
It defaults to the Smart tier because reading a pipeline and judging which deals are at risk benefits from a stronger model. You can change the tier per routine to control spend.
Where does the summary go?
To your local output history on your Mac, one rich card per run. Extend the prompt to also post it to Slack or email it to yourself.
More Business recipes
Keep exploring
Competitor Watch
A Monday-morning intel brief on your competitors: new launches, pricing changes, and notable announcements, with what each move means for you.
Daily Jira Summary to Slack
A weekday-morning Slack post that summarizes what changed in your Jira project overnight: issues moved, blockers, and what is due today.
Candidate Screening Brief
A weekday brief on new candidates from your hiring inbox: a fit read against the role, strengths and gaps, and tailored questions to ask in the first call.
Weekly Investor Update Draft
A Friday-afternoon investor update drafted from your week, highlights, key metrics, lowlights, and your specific asks, in a clean format ready to review and send.
Morning Briefing
A single written brief waiting for you at 7:30 AM: today's calendar with a prep note on each meeting, your email split into reply now and can wait, and the commitments you made that are still open.
Email Digest
A ranked digest of overnight unread mail, the must-reply messages at the top, the newsletters and noise summarized at the bottom, ready when you sit down.
Run Daily CRM Pipeline Summary on your Mac.
Install Routines, paste the prompt, set the schedule. It runs locally and the output stays on your machine.
Step 01 · Install Routines. Download the DMG, drag to Applications, and sign in with Google once.