Skip to content
agents/en vrai

Get alerted when a client has waited too long for your reply

Every hour, the agent finds client emails in Gmail you haven't answered yet, has the AI judge whether a reply is really needed and pings you on Slack as soon as your target response time is reached.

L3 autonomousVerified sourceBased on official documentation we checked; not tested by us yet.Updated
Difficulty
Intermediate
Setup time
≈ 1 h 30 (estimate)
Estimated cost
n8n: €0 self-hosted or from €20/month (Starter, billed annually) + OpenAI API calls; Slack: free plan available
Category
Small business & freelancers

⏱ Trigger

Every hour (cron)

✋ Human approval

Read-only on Gmail: the agent never writes or sends an email. You do the replying.

→ Output

Slack alerts (due, then overdue) + Google Sheets log of your response times

Autonomy level 3 — Autonomous: Runs end to end on its own; only low-risk actions (read, notify, archive).

01Connected tools

02What the agent does on its own

  1. step 01

    Every hour, reads recent Inbox and Sent messages and matches them by thread to find client emails still waiting for your reply.

  2. step 02

    Checks them against a Google Sheets log to separate new cases, overdue ones and threads you have already answered.

  3. step 03

    For each new case, has OpenAI decide whether a reply is expected, with a priority and a one-sentence summary.

  4. step 04

    Logs the thread and posts a Slack alert only when a reply is needed; escalates once the wait passes twice your target time.

  5. step 05

    Once you reply, records when you did and your response time in hours.

03Agent instructions

$ instructions.txt · replace the {VARIABLES}
> Client email from {SENDER} on {DATE}, subject "{SUBJECT}": "{EXCERPT}". Is a reply from me expected? Return JSON {"reply_needed":true|false,"priority":"high|normal|low","summary":"one sentence"}. A simple thank-you or acknowledgement does not need a reply.

04Limits and legal points

One run per hour is roughly 720 executions a month: check your n8n plan's execution allowance. Exclude your own domains and list your clients' domains, or newsletters and notifications will look like clients. Replies made by phone or from another address are invisible to the agent. Email excerpts are sent to OpenAI (GDPR). To be alerted somewhere other than Slack, you'll need to change the alert steps.

05In short

When does the agent run, and what does it produce?
Trigger: Every hour (cron). Output: Slack alerts (due, then overdue) + Google Sheets log of your response times.
What stays under your control?
Read-only on Gmail: the agent never writes or sends an email. You do the replying. Autonomy level 3 — Autonomous: Runs end to end on its own; only low-risk actions (read, notify, archive).
How much does it cost and how long does setup take?
Estimated cost: n8n: €0 self-hosted or from €20/month (Starter, billed annually) + OpenAI API calls; Slack: free plan available. Setup time: ≈ 1 h 30 (estimate), difficulty "intermediate". Tools: n8n – Schedule Trigger, n8n – Gmail, n8n – OpenAI, n8n – Google Sheets, n8n – Slack.

Answers taken from this recipe · updated 22 October 2026

06Source

https://n8n.io/workflows/21255-track-client-reply-slas-in-gmail-with-openai-google-sheets-and-slack/

07Related recipes

Same category or same connected tools.

All recipes in "Small business & freelancers" →

→ weekly newsletter

3 agents en vrai.
Every week.

Three reproducible, sourced autonomous agents. No hype, no "top 100 tools". One-click unsubscribe.