Example 1: signup → saved subscriber → welcome email
AutomationPicks implements this sequence in its deals signup action. The steps below describe the current code, not a measured delivery rate or a confirmed email receipt. Saving a subscriber and delivering an email are separate outcomes.
- Trigger and inputs: a signup submits email and source, with an optional page_url. The saved record also gets the automation-deals tag.
- Step 1 — validate before writing: trim and lowercase the email, reject invalid addresses, and keep it within 254 characters. The action caps source at 64 characters and page_url at 500.
- Step 2 — save the subscriber: insert the record before attempting email delivery. A storage error other than an existing subscriber stops the signup.
- Step 3 — handle repeats: an already-subscribed address returns success without sending another welcome. A new address triggers the welcome-deals template with a per-address idempotency key.
- Failure boundary: if welcome sending fails, the saved subscriber remains and signup still succeeds. The action logs the send failure; it does not implement a retry queue or prove delivery to an inbox.
- Before using this pattern: submit a new test address, check that one record is saved, then submit it again and confirm it does not trigger another welcome. Check email delivery separately rather than treating the signup success message as a receipt.
Example 2: vendor pricing check → validation → change record
The site's pricing-refresh implementation fetches configured vendor pricing pages and extracts structured plan data. It compares the result with the previous stored snapshot. This is a pricing check, not a partner-deal feed, and it cannot guarantee that a vendor page is complete or that every published claim is updated.
- Required configuration: each vendor needs a name, tool slug, pricing-page URL and the list of affected page paths. The extracted snapshot includes currency, free-plan availability and plan name, price, display text and billing period.
- Step 1 — fetch and extract: strip scripts and markup from the vendor page, then request structured pricing. Failed fetches, unclear pricing or an invalid extraction are reported as failed checks instead of being replaced with guessed prices.
- Step 2 — compare like-for-like: use the lowest numeric paid entry price and compare it with the saved snapshot. Keep billing period and currency in view when reviewing the result.
- Step 3 — hold suspicious changes: an entry-price move greater than 40% in either direction is flagged for review, and its history record stays unpublished. This guard is implemented; it is not evidence that smaller changes are always accurate.
- Step 4 — record an accepted change: save the snapshot and check time. When an existing entry-price display changes, add a history record and mark the configured pricing pages as updated.
- Limitations: blocked pages and ambiguous annual-versus-monthly pricing can prevent extraction. A plan-detail change without an entry-price display change does not use the same page-update path. A changed date is not proof that every sentence on a review was rewritten.
- Acceptance check: review the extracted plan against the vendor page, compare old and new records, and read the affected public page. Treat flagged and failed checks as work to review, not successful refreshes.
Workflow ideas versus evidence of a run
The examples below are build ideas, not a claim that AutomationPicks has run each one. For a real test, record the trigger, input fields, destination record, runtime, usage and failure details. Do not publish draft run data as measured results.
- Confirm the downstream record or publish result, not just the trigger firing or a green step indicator.
- Test missing required fields, a repeated event and an unavailable destination before relying on the workflow.
- Measure usage from the platform's run history. A multi-step run can consume more than one task or credit, so workflow count alone cannot determine cost.
- Keep credentials and customer data out of public run notes. Describe the fields and outcome without publishing the underlying personal records.
Lead capture and sales workflows
These are the highest-ROI automations to build first: every one of them shortens the gap between someone raising their hand and someone following up.
- New form submission → create CRM contact → notify sales in Slack (Fillout + Make)
- Qualified lead → book a call link by email; unqualified → nurture sequence (Make router)
- New Stripe payment → tag the buyer in your email tool → start onboarding emails (Pabbly Connect)
- Abandoned checkout → wait 2 hours → send a recovery email (Systeme.io)
- Inbound webhook from your website → enrich the record → route by deal size (Make)
Client onboarding and delivery workflows
Agencies lose the most time here, because onboarding is repetitive but never quite identical. Automate the repeatable 80% and leave a manual approval step for the rest.
- Signed proposal → create project folder, task template, and shared doc (Make)
- Intake form submitted → generate a welcome PDF → email it with next steps (Fillout + Pabbly)
- Kickoff date set → schedule reminder emails at T-7, T-2, and T-0 days
- Deliverable marked complete → notify the client and request feedback
Finance and admin workflows
Money workflows are where automation quietly pays for itself. Each of these removes a recurring manual task that is easy to forget and expensive to miss.
- New invoice paid → update the revenue sheet → post the total in Slack (Pabbly Connect)
- Invoice unpaid after 7 days → send a polite reminder, then escalate at 14 days
- Receipt email received → extract the amount → file it in your bookkeeping tool
- Month end → compile revenue, refunds, and MRR into one summary email
Content, marketing and reporting workflows
These keep publishing and reporting consistent without adding a person to the process.
- New blog post published → cross-post to your newsletter and social queue
- Weekly analytics pull → build a client report → email the PDF every Monday (Make)
- New review or testimonial submitted → route to approval → publish to the site
- New subscriber → tag by source → drop into the matching email sequence (Systeme.io)
Which tool fits which example
Most of these can be built on any platform, but the cost and effort differ a lot depending on the shape of the workflow.
- Branching, iterators, and multi-app logic — Make handles complexity best
- High run volume where per-task pricing hurts — Pabbly Connect stays flat as you scale
- Form-driven workflows with conditional logic — Fillout is the cleanest front door
- Funnels, email sequences, and checkout flows — Systeme.io keeps it in one place
How to adapt any example to your business
Take an example above, then map it onto your own stack in four passes: name the trigger, list the data you need, write the actions in order, and add one filter that stops bad data early. Build the happy path first, run it against a test record, then add error handling before you point it at production.
