Personalising WhatsApp auto-replies for new versus returning customers is the difference between a generic bot experience and a conversation that feels like it recognises the person on the other end. Stitch, the WhatsApp business communication platform, is built on keyword-triggered automation and flow design that make this kind of segmentation possible without asking customers to repeat themselves every time they message in.
This guide walks through how to configure auto-replies so new contacts get a proper introduction and qualification journey, while returning contacts skip straight past the menu they've already seen.
Before you start, have the following ready:
Access to your WhatsApp automation settings, including default replies, delay timers and flow builder.
A clear view of your existing tags or contact segments, since tagging is what separates "new" from "returning" in most setups.
Agreement internally on what "returning" means for your business (e.g. anyone with a prior resolved chat, anyone tagged as a customer, anyone with an active CRM record).
A defined human-takeover rule so the bot doesn't re-trigger once a staff member is already handling the conversation.
A configurable default reply acts as a catch-all when an inbound message doesn't match a programmed keyword or phrase, and it's the foundation for any new-versus-returning logic. Alongside this, default-reply delay and away-reply delay settings control when a customer sees the opening flow again, and automated-reply duration determines how long an unfinished flow stays active while waiting on a response. Together, these settings let you decide not just what a new or returning contact sees, but when they see it.
Human-takeover duration is equally important here: once a staff member replies, the bot should stop interrupting that conversation, which matters for returning customers who are mid-conversation with an agent rather than starting fresh.
Keyword-driven automations can still run where their configured trigger matches, even alongside a default catch-all reply, which is what allows a business to route new enquiries down one path and existing relationships down another. During onboarding, this was demonstrated by directing customers to select from a menu of options such as "new inquiry," "existing inquiry" or "existing booking," with each selection triggering an exact phrase match that initiates a tailored conversation flow. This structured approach reduces ambiguity and speeds up data collection compared to open, free-text replies.
For a new-versus-returning setup, mirror this logic:
Build a first branch for contacts with no prior tag or CRM record, and route it into a full introduction and qualification flow.
Build a second branch for tagged or previously-engaged contacts, and route it straight to "how can we help today" or a shortcut back to their existing team or agent.
Individual contacts can have automations disabled entirely when an ongoing relationship makes repeated menus inappropriate. This is the cleanest way to handle long-standing customers, key accounts, or anyone who has an established point of contact: rather than trying to write a "returning customer" message that fits every scenario, you simply switch off the catch-all for that contact and let the conversation flow directly to the team or person already assigned.
This matters operationally too. After an outbound template or bulk send, the default response can be suppressed so that a meaningful customer reply doesn't trigger an irrelevant welcome menu. If a returning customer replies to a broadcast or reminder, the last thing you want is a "Hi, welcome, are you new here?" flow undoing the personalisation you've already built.
New contacts benefit most from a flow that asks structured questions and collects responses before assigning the conversation to the appropriate person or team. A well-designed new-customer journey typically:
Sends an approved opening message introducing the business.
Asks a small number of qualifying questions relevant to the sector (for example, sales or lettings for a property business, or service versus sales for automotive). For property agents, this is a highly effective way to respond to Rightmove leads faster using WhatsApp automation by capturing requirements immediately.
Routes the conversation to the correct team once the necessary information has been captured, rather than assigning it to whichever staff member happens to be free.
Teams are best kept assigned to a function (such as bookings or lettings) rather than auto-assigned to the last sender, which centralises responsibility and avoids a chat sitting with someone who is unavailable. Manual reassignment then gives staff the flexibility to pick up conversations as needed.
For returning or dormant contacts, the priority is different: skip the introduction and confirm intent quickly. Where a reviewed, consent-appropriate segment has been selected, send a relevant approved message with simple choices such as "still interested," "update details" or "stop receiving messages". This respects unsubscribe status and account quality, and updates the CRM only where the integration supports the change.
For contacts re-engaging after a call rather than a previous chat, a similar principle applies: if the caller chooses to continue on WhatsApp, send the approved opening message, ask why they contacted the business, then route to the relevant person or team. This keeps the tone appropriate to someone who already has a relationship with the business, without repeating a full new-customer script.
Whichever branch a contact lands in, be explicit about response times. Communicating a realistic response time, such as "within two working hours," inside the auto-reply manages customer expectations and improves satisfaction. This is worth including in both the new-customer and returning-customer paths, since new contacts in particular need reassurance that a human will follow up.
Any personalisation strategy has to work within WhatsApp's messaging rules. Template messages are required to restart conversations outside the 24-hour window, which protects the business from policy violations, and these templates can be personalised and include quick-reply buttons to keep engagement high even when re-opening a lapsed conversation. This is especially relevant for returning or dormant customers, where the gap since the last message is more likely to exceed 24 hours.
Once both flows are live, confirm the bot correctly steps aside once a person takes over. The bot should be "out of sight, out of mind," automatically disabling when a human agent replies, and only resuming if the customer messages again after the agent stops typing. Test this for both new and returning journeys, since a returning customer who's been shortcut straight to an agent is the group most likely to notice (and be frustrated by) a bot re-inserting itself.
Tag early, tag consistently. The new-versus-returning distinction only works if contacts are tagged accurately at the point of first contact, since broadcasting and automation both rely on tagged contacts to target the right group.
Keep menus exact-match, not free-text. Avoiding free-text inputs in favour of keyword-triggered, exact-phrase paths makes both new and returning journeys easier to manage and automate.
Archive resolved chats. Archiving chats after resolution keeps the inbox organised and makes it easier to see, at a glance, which contacts are genuinely new versus simply inactive.
Don't present WhatsApp as a replacement for every voice interaction. Avoid presenting a WhatsApp journey as a guaranteed replacement for every voice conversation or emergency call, particularly when designing the new-customer intake flow.
Remember the channel preference behind this effort. Personalising the experience matters because a majority of consumers prefer reaching businesses via WhatsApp over email and telephone, so a clumsy, one-size-fits-all auto-reply undermines the very reason customers chose the channel.
A returning customer keeps getting the new-customer menu. Check whether their automations have been disabled at the contact level, and confirm the default-reply delay hasn't reset. If they're replying to a broadcast or template send, verify the default response suppression setting is active for that send.
The bot interrupts a human agent mid-conversation. Review the human-takeover duration setting, which is designed specifically to prevent the bot from interrupting an active conversation once a staff member has replied.
New contacts aren't being routed to the right team. Confirm the flow is asking structured questions and capturing responses before assignment, rather than assigning on first message, and check that team-based assignment (not last-sender assignment) is configured.
Messages to dormant contacts aren't being answered as expected. Confirm the correct customer or candidate segment has been reviewed for consent before sending, and check that the message includes simple choices rather than an open-ended prompt.
Personalising WhatsApp auto-replies for new versus returning customers comes down to a handful of configurable mechanics already built into Stitch: default replies and delays, keyword-triggered exact-match flows, per-contact automation controls, and team-based assignment logic. Get the new-customer flow to ask the right qualifying questions before handover, get the returning-customer flow to skip straight to intent and routing, and make sure the bot steps back the moment a human takes over. Configure these correctly once, and every customer, whether it's their first message or their fiftieth, gets a conversation that fits where they actually are.
