WhatsApp Auto-Reply Rules: How Keyword Automation Actually Works

An auto-reply system is not a chatbot. It does not understand the message — it matches a pattern and sends a response. Every auto-reply rule is three parts: a trigger that says when to consider replying, conditions that narrow when it is appropriate, and the response itself.

Most auto-replies that annoy people are missing the middle part. This article covers all three, plus the ordering and rate-limiting behaviour that determines whether the system helps or embarrasses you.

The Three Parts of a Rule

1. The Trigger

The trigger decides which incoming events are candidates. Common trigger types:

Trigger	Fires when
Keyword contains	The message includes any of your listed words
Exact match	The message equals a specific string, such as a menu number
Starts with	The message begins with a prefix, useful for command-style input
Any message	Every incoming message, used for away messages
New chat	The first message ever received from that number
Regular expression	Pattern matching for order numbers, dates, structured codes

Keyword triggers should be lists, not single words. People do not type the word you expect. A pricing rule needs price, pricing, cost, how much, rate, rates, quote, charge at minimum, and it should match case-insensitively.

2. The Conditions

Conditions are what separate a helpful automation from an irritating one. A trigger says “this message mentions price.” A condition says “…and it is 11pm, and this person has not received this reply already today, and no human has spoken to them yet.”

Conditions worth having:

• Time of day and day of week. Away messages should not fire during business hours.
• Contact type. Saved contacts, unsaved numbers, or a specific list. Existing customers rarely want the new-enquiry autoresponder.
• Chat type. Individual chats versus groups. Auto-replying in groups is almost always wrong and is the single fastest way to make automation look broken to a room full of people.
• Cooldown. Do not fire this rule for this contact again within N hours.
• Human-handled suppression. Once you reply manually in a thread, stop automating it.

The last two are not optional. Without a cooldown, a customer who sends three messages in a row gets three identical replies, which reads as malfunction rather than service.

3. The Response

The response is the template that goes out. Good responses share four properties:

• They disclose the automation. “This is an automatic reply” costs one line and buys enormous goodwill.
• They set a time expectation. “Someone will reply by 10am tomorrow” is far better than silence about timing.
• They provide an escape hatch. A phone number, an email, or “reply URGENT and we will be notified.”
• They answer something. An auto-reply that only says “we got your message” wastes the interaction. Include the business hours, the price range, the link — whatever the trigger implied they wanted.

Rule Ordering Matters

Rules are evaluated top to bottom, and in most implementations the first match wins. That makes ordering a functional decision, not a cosmetic one.

Order from most specific to most general. If your “any message” away rule sits above your keyword rules, it will match everything and none of the keyword rules will ever fire.

A working order:

1. Opt-out handling (stop, unsubscribe) — always first, always wins
2. Exact-match menu selections (1, 2, 3)
3. Specific keyword rules (pricing, hours, location, booking)
4. New-chat greeting
5. Out-of-hours away message
6. Catch-all fallback, if you want one at all

Opt-out first is non-negotiable. Someone typing “STOP” must never receive a marketing autoresponder in return.

A Realistic Rule Set

Most businesses need five or six rules, not fifty.

Rule 1 — Opt-out Trigger: exact match on stop, unsubscribe, remove me. Condition: none. Response: confirm removal, and flag the contact so no future automation or campaign includes them.

Rule 2 — Business hours Trigger: keywords hours, open, closed, when are you open. Response: your hours, plus today’s status if you can express it.

Rule 3 — Pricing Trigger: keywords price, cost, how much, quote, rates. Condition: individual chats only. Response: price range or starting price, plus what determines the final figure. Attach the price list. Never reply “it depends” and nothing else.

Rule 4 — Location and directions Trigger: keywords where, address, location, directions, parking. Response: address, a map link, and the one landmark detail people always ask about.

Rule 5 — Out of hours Trigger: any message. Conditions: outside business hours, individual chats only, 8-hour per-contact cooldown, not already handled by a human. Response: acknowledgement, next opening time, and an urgent escape hatch.

Rule 6 — New chat greeting Trigger: first message from a number. Condition: during business hours only, so it does not collide with Rule 5. Response: welcome, what you do, and the one question you need answered to help them.

That is a complete system. It handles the majority of inbound volume and it is small enough to audit in five minutes.

What Auto-Replies Should Not Do

• Do not reply in groups. Almost never correct, and highly visible when wrong.
• Do not chain. Two automated systems replying to each other is a real failure mode. Never auto-reply to a message that itself looks automated.
• Do not pretend to be human. People find out, and it costs more trust than the automation saved.
• Do not handle complaints. An angry message deserves a person. Route it, do not answer it.
• Do not fire on every message in a conversation. One well-timed automated message is service; four is noise.

The Availability Constraint

Auto-replies that run in a browser extension execute inside your open WhatsApp Web session. Close the tab and the rules stop. Sleep the laptop and they stop. This is a genuine limitation and pretending otherwise leads to customers who were promised a reply and got nothing.

Two ways to work with it honestly:

• Keep a dedicated browser window open on a machine that stays awake during the hours you have promised coverage.
• Word the away message so it does not over-promise: state when someone will reply, not that a reply is guaranteed instantly.

WAPlus evaluates auto-reply rules locally in the page while WhatsApp Web is open — no message content is sent to a server for matching — which is the right privacy trade-off, and it is also why the tab has to stay open.

Measuring Whether It Works

Track three things per rule:

• Fire count. A rule that never fires should be deleted or its keywords widened.
• Human-takeover rate. How often a person still has to step in after the automation replied. High rates mean the response is not actually answering.
• Complaint rate. Any “stop sending me this” is a design bug in a rule, not user error.

Review monthly, delete what does not fire, and rewrite what always triggers a takeover.

Related reading: WhatsApp Message Templates covers writing the responses your rules will send.