Automated follow-up workflows: 5 n8n templates (quotes, invoices, files, carts)

A quote that never got a reply, an invoice past its due date, a case file stuck over one missing document, a cart abandoned at checkout: every one of these scenarios has the same remedy — a follow-up — and the same failure point — nobody has time to do it methodically. That's exactly why the follow-up workflow is automation's number one use case: a perfectly repetitive task, low value per unit but with a high cost of forgetting, where the machine beats the human on the only criterion that matters — consistency.

What makes the topic even more interesting is that the effectiveness of reminders is solidly documented. A field experiment by Ximena Cadena and Antoinette Schoar published in 2011 (“Remembering to Pay? Reminders vs. Financial Incentives for Loan Payments”, NBER) shows that simple SMS reminders improve on-time loan repayment about as much as costly financial incentives — a message at the right moment is literally worth a discount. In the same vein, the study by Karlan, McConnell, Mullainathan and Zinman published in Management Science in 2016 (“Getting to the Top of Mind: How Reminders Increase Saving”) demonstrates that reminders significantly increase the desired behavior simply by bringing the deadline back to mind. In other words: most ignored quotes and overdue invoices aren't bad faith, they're forgetfulness — and forgetfulness is fixed with a workflow.

This article is a hub guide: first the anatomy common to every follow-up workflow in n8n, then five concrete templates, each pointing to its dedicated deep-dive.

The anatomy of a follow-up workflow

Whatever you're chasing — quote, invoice, case file, cart, contract — the skeleton is always the same:

Schedule Trigger (daily, 9am)
        │
        ▼
Source of truth (CRM / Sheets / database)
        │  pending items
        ▼
Age calculation (day 3 / day 7 / day 15)
        │
        ▼
Stop condition? ── replied / paid / completed ──► exit the sequence
        │ no
        ▼
Escalation tier (soft email → firm email → SMS → human)
        │
        ▼
Drafting (AI) + send
        │
        ▼
Logging (date + reminder counter)

The source of truth

Everything starts from a single place where the items to follow up and their statuses live: a CRM (HubSpot, Pipedrive), a Google Sheet, a Postgres or Supabase database. That's what the workflow queries every morning, and that source — never the workflow itself — decides whether an item is still pending. If your customer data is scattered across tools, start by syncing your CRM with the rest of your stack, otherwise your follow-ups will fire on stale statuses.

The scheduled trigger

A daily Schedule Trigger (typically 0 9 * * 1-5: every weekday morning at 9am) kicks off the run. The setup has more subtleties than it looks — cron expression, business days, and above all the server's timezone — all covered in our Schedule Trigger and timezones guide.

The age calculation

For each pending item, the workflow computes how many days have passed since the reference event (quote sent, invoice due date, file opened). That number determines the tier: nothing before day 3, first follow-up at day 3, second at day 7, last at day 15. An IF or Switch node then routes each item to the right tier.

The stop conditions

This is the most important part of the workflow, and the most often botched. Before every send, the workflow verifies the item still needs a follow-up: quote unsigned, invoice unpaid, file incomplete, cart not converted. A customer who receives a chaser for an invoice they settled yesterday instantly loses trust in everything you send afterwards. The stop condition is checked at send time, against the live status in the source of truth — not against a snapshot taken the day before.

Progressive escalation

A good sequence increases in intensity without ever turning aggressive: a soft first email ("in case this slipped past you"), a more direct second one, then a channel switch — SMS or a call — and finally an escalation to a human, who takes over with the full history in front of them. Automation handles tiers 1 through 3; tier 4 is an internal notification, not a customer message.

Logging

Every follow-up sent gets logged: last-send date (last_reminder_at) and a reminder counter, written to the source of truth right after the send actually succeeds. That's what guarantees an item is never chased twice the same day, that tiers fire in order, and that a human opening the record sees exactly what has already gone out.

AI personalization

The top layer: instead of a fill-in-the-blank template, an LLM drafts each message from the real context (name, subject, age, reminder number). Two recipients never receive the same wording, and the tone naturally firms up with the counter. The mechanics are the same as for AI-generated email reply drafts: context injected into the prompt, tone guidelines, output ready to send.

Template 1: following up on unanswered quotes (CRM)

The quote is the textbook case: a salesperson sends a proposal, the prospect goes quiet, and without a follow-up the deal dies in silence. Every morning the workflow queries the CRM for deals at the "quote sent" stage, computes the age, and runs the sequence — day 3 (light nudge with the quote link), day 7 (offer of a call), day 15 (a closing message that, paradoxically, often triggers a reply). Stop condition: the deal changed stage. Clean CRM data is the absolute prerequisite — a well-synced pipeline, as described in our guide on syncing HubSpot or Pipedrive with n8n, keeps you from chasing a prospect the salesperson already had on the phone.

Template 2: chasing unpaid invoices (Stripe)

Invoice follow-up is the one that translates most directly into revenue. With Stripe, the workflow can even be event-driven rather than scheduled: a webhook signals the failed payment or overdue invoice, and n8n runs the sequence — a courteous reminder with the payment link, a firmer second one, then an internal notification for human handling. Every message carries an up-to-date payment link, and the stop condition is verified against Stripe itself: invoice paid, sequence over. The full build, webhook included, is detailed in payment follow-ups with Stripe, webhooks and AI in n8n.

Template 3: chasing incomplete case files and missing documents

HR, compliance, insurance, training: wherever supporting documents are collected, a large share of files stalls for lack of active follow-up. Each morning the workflow lists the incomplete files in the database, has the AI draft an email that names precisely the missing documents (and only those), sends, logs, and posts a run summary to Slack. It's the most detailed template of the series: our guide on automating follow-ups for incomplete case files with n8n builds it node by node, from the Supabase table to the Slack summary.

Template 4: recovering abandoned carts (e-commerce)

The abandoned cart is the only template where the clock runs in hours, not days: a follow-up at hour 1 or hour 3 with the cart contents, possibly a second at hour 24. AI personalization makes a real difference here — a message mentioning the exact products in the cart converts better than a generic "you forgot something". Obvious stop condition: the order went through. The complete workflow, from abandonment detection to the generated email, is in recovering abandoned carts with AI and n8n.

Template 5: reminders ahead of contract deadlines

The last template, and the only one that follows up before the event rather than after: subscriptions, maintenance contracts, certifications up for renewal. The workflow scans deadlines at day −60, −30 and −7 and alerts the right person — the customer for a renewal, the internal team for a renegotiation. Same skeleton, inverted logic: here, logging prevents alerting three times for the same deadline. The setup is described in tracking contract deadlines with n8n.

Email or SMS: which channel for which follow-up?

Email is the default channel: unobtrusive, it carries a substantive message, a payment link, an attachment, and leaves a written trail. Its weakness is well known — it drowns. SMS, by contrast, almost always gets read, but its intrusiveness reserves it for two uses: genuine urgency (a badly overdue invoice, an imminent deadline) and escalation after two unanswered emails. So the right practice isn't choosing, it's sequencing: email for tiers 1 and 2, SMS for tier 3. Wiring Twilio into n8n, costs and best practices included, is covered in sending SMS with Twilio and n8n.

The classic mistakes

• Following up without a stop condition. The cardinal sin: a chaser sent to someone who already paid, replied, or completed destroys the credibility of the entire sequence. Check the status at send time, not the day before.
• Skipping the logging. Without last_reminder_at and a counter, the same recipient can get two follow-ups on the same day — or the sequence can restart from scratch after a workflow restart.
• Pushing too hard, too fast. Three follow-ups in five days for a B2B quote is harassment, not diligence. Match the cadence to the object: hours for a cart, days for a file, weeks for a contract.
• Ignoring timezones and send windows. A follow-up that lands at 3am or on a Sunday instantly gives the automation away. Set the Schedule Trigger timezone and keep sends within business hours — the cron and timezones guide covers exactly this.
• Automating the final tier. The last follow-up — the one before formal notice or closing the file — deserves a human. The workflow prepares the context and notifies; it doesn't replace judgment.

Where to start

Pick the template where forgetting costs you the most — for most teams, that's the unpaid invoice or the unanswered quote — and build that one alone, end to end: source of truth, trigger, age calculation, stop condition, logging. Once that first workflow is in production, the other four come together by reusing the same skeleton and simply swapping the source, the timings, and the tone. Reminders are among the best-documented interventions in behavioral economics; it would be a shame if your business were the only one not cashing in on them.