How to Build a Missed-Call Text-Back Workflow in your CRM

Start with the real missed-call event

A missed call is not always a missed lead. The same person may call twice, leave voicemail, or speak with the team after the first attempt. Before you automate a reply in your CRM, decide exactly which inbound calls should count. Is the number dedicated to new sales enquiries, or does it also receive existing-customer support? Should an unanswered call outside office hours be handled differently from one missed at lunchtime? Make those decisions for the actual sub-account, not for a generic agency template. Record the number, the business name callers expect to hear, the hours a human can respond, and the person responsible for the conversation. A useful outcome is a single, timely acknowledgment followed by a real path to help; sending a text for its own sake is not the goal.

Check the built-in option before making a workflow

CRM already has a built-in Missed Call Text Back setting. Its own documentation warns that it can send a message for every missed call, including repeated calls within a short period. If that behaviour is acceptable, you may not need a custom workflow at all. If you need a cooldown, different messages by time of day, or a team handoff, a workflow gives you a place to design and test those conditions. Avoid leaving the built-in setting and a new workflow active for the same number without checking their interaction: two independent replies make the customer experience worse. Ask the account owner whether a missed-call reply already exists, inspect the current settings and workflows, and agree who owns the final message before creating another automation.

Choose a trigger and filter deliberately

For a custom workflow, use the call-related trigger available in the account and narrow it to the relevant inbound number and unanswered outcome. CRM documents a Call Details workflow trigger with filters for call direction and call status. Verify the exact status names in your account: “not answered,” voicemail, busy, and canceled can mean different things operationally. Do not substitute a broad “contact created” trigger, which may also fire for imports and form submissions. Decide whether voicemail deserves a text or whether someone will respond manually. A test with the right trigger should enroll the intended caller; a test answered call and an unrelated number should not. Write down the entry rule in one sentence so another teammate can verify it without reading the entire workflow.

Design a cooldown without losing real enquiries

Repeated calls are normal when somebody is trying to reach a business. A workflow that replies on every attempt can generate several almost identical texts. Choose a documented cooldown rule for your operation, such as one acknowledgment per contact and number during a short window, and make sure a legitimate new enquiry later can still get attention. The exact duration is a business decision; do not copy a number from somebody else’s workflow blindly. A wait step can provide time for a person to call back, but a wait alone does not prevent two separate enrollments from sending twice. Check the workflow’s re-entry settings and any tag or condition you use as a duplicate guard. Test successive calls from the same number and a call from a different number to see how the actual account behaves.

Get messaging permission and identity right

An unanswered call does not automatically grant permission for every kind of promotional text. Check the sub-account’s messaging setup and the consent basis that applies to this particular service, jurisdiction, and contact. If the contact has opted out or a number cannot receive texts, do not attempt a workaround. Identify the business in the message so it is not an anonymous surprise, and keep the first reply useful rather than promotional. A practical draft might say: “Hi, this is [business]. Sorry we missed your call. What can we help you with? Reply here and our team will follow up during [hours].” Replace every bracket with a real, verified detail before publishing. If hours or a callback promise are uncertain, remove the promise. Ask the client to approve both the message and the channel before a live test.

Handle business hours and ownership

An automated acknowledgment is not the same thing as a human response. If the team only answers during set hours, say so and set the expectation honestly. An after-hours call can enter the same workflow but may need different wording and a task for the next working day. Assign a clear owner for inbound replies and decide what should happen if that person is unavailable. An internal notification should include enough context for a teammate to find the conversation, without sending contact details to the wrong sub-account. If your agency manages several clients, use each client’s own phone number, operating hours, staff assignment, and approved copy. One account’s identity or calendar link must never appear in another account’s message.

Stop when a human conversation begins

The follow-up should change when the person replies, books, or speaks with the team. CRM documents a workflow-level Stop on Response setting, but that option has specific behavior in workflows that also use call actions; review its current documentation and test it rather than assuming every reply or voicemail ends every path as intended. Decide whether a staff takeover should stop scheduled texts too. If you add a second reminder, confirm the contact has not already replied, opted out, booked, or been contacted manually before it sends. Keep the initial design simple: one acknowledgment and a clear team handoff are easier to verify than a long sequence. If a branch cannot reach the right next step, the workflow may look complete while leads quietly go unanswered.

Build a draft in Lead Spine

In Lead Spine, connect and choose the CRM sub-account that owns the phone number. The recommended connection path is the marketplace app; some workflow-builder actions can also require your authorized agency sign-in. Then give Claude, ChatGPT, or another connected assistant a specific brief: “For [named sub-account], draft a missed inbound call follow-up for [number]. Exclude answered calls, respect opted-out contacts, avoid duplicate replies, use [approved message], and create a human follow-up task for [team]. Do not publish it.” Replace the placeholders with real account details; never put passwords or verification codes into the conversation. Ask the assistant to summarize the trigger, filters, wait, stop conditions, message, and target account before it builds. If a capability or approved access is unavailable, stop and resolve the setup instead of accepting a guessed result.

Review the draft against the account

Inspect the resulting draft in CRM itself. Verify the target sub-account and number first, then follow the path from trigger to last action. Is the call direction inbound? Does the condition match an unanswered call rather than any completed call? Does the duplicate guard run before the send step? Check the sending number, the staff recipient, each custom field, and any booking link. Check whether another workflow or the built-in setting handles the same event. If the assistant reports a successful build, that is a prompt to inspect the actual draft, not proof that it is safe to publish. Save a short review note recording the intended result and any assumptions the client still needs to approve.

Run a useful test matrix

Test with safe numbers and consented test contacts, not a live lead you cannot control. Try one missed inbound call during business hours, then a second call from the same number in quick succession. Try an answered call, a voicemail if you support it, an after-hours missed call, and a call to an unrelated number. Reply to the text, then try booking or handing the conversation to a human. Confirm how opt-out and an undeliverable number behave. For each case, compare the contact’s timeline and workflow activity with the expected outcome: was the contact enrolled, was a text sent once, and did the follow-up stop at the right moment? A failed test is valuable; adjust the draft and repeat before you enable it.

Publish with a monitoring plan

After the business owner approves the copy and the test results, publish the workflow intentionally. For the first days, watch a small sample of real call and conversation timelines for duplicate sends, wrong-hour promises, delivery failures, and replies without a team response. Keep a record of what was enabled and when so another operator can reverse it. If you change a phone number, office hours, or assignment rules later, re-test the affected paths. Review the built-in setting whenever you migrate to or from a custom workflow so one missed call does not receive both replies. Success is not the number of automations launched; it is whether the caller got an appropriate response and the team can continue the conversation.

Common failure: “the text sent, so it worked”

The first text is only the beginning. An automated message may send successfully while the next action points to an inactive user, the booking link belongs to a different calendar, or the team never notices a reply. Check both delivery and handoff. Ask the client where a reply appears and who is responsible for it; if nobody owns that inbox, fix ownership before scheduling more messages. If the contact has already spoken with a person, a later automated reminder can sound careless. Follow the complete conversation, not only the send event. If the account is shared between teams, verify that assignments and notifications do not expose conversations to the wrong people.

Common failure: a workflow is published but never starts

If no test contact enters the workflow, revisit the trigger before changing message copy. Confirm the sub-account that received the call, the number to which it was routed, and whether the provider classified the outcome as missed, voicemail, busy, or another call status. Compare the event against each filter and the enrollment or re-entry rule. If a test call does not create the event you assumed, changing a wait step will not help. Conversely, if an answered call enters, the trigger is too broad. Test an intended event and a deliberate near miss whenever you change call routing or account configuration. Leave the workflow in draft until the correct events enroll reliably.

Common failure: consent or opt-out is treated as a footnote

Operational convenience does not override the contact’s preferences or applicable messaging requirements. Review how the client collects permission to text, how opt-outs are stored, and whether the message type is permitted in the places they serve. Check this before activating sends, particularly when copying a workflow from another client. Keep an audit trail of the message that was approved and the account in which it was sent. A caller who has opted out should not receive an automated response just because a technical branch is easy to build. If the business is unsure about the legal basis, pause the send and get qualified advice rather than inventing consent rules.

What to do next

Start with one sub-account and one agreed missed-call use case. Inspect the existing built-in setting, write the trigger and message in plain language, confirm messaging permission, and draft the workflow. Test the five key cases—missed, answered, repeated, replied, and opted out—before publishing. If an AI assistant is doing the setup, give it the exact account and require a draft and a readable explanation. You can use Lead Spine to connect an assistant and review the account, or you can build directly in your CRM. In either case, use the actual call and conversation timelines to decide whether the automation is helping; a workflow is valuable only if it produces the right experience for the person who called.