The Missed-Call Text-Back and After-Hours Intake Playbook

A service business that does not answer is usually a business that is busy doing the work, which is the reason the calls pile up after hours and in the middle of jobs. A caller who reaches voicemail often moves on to the next listing. The text-back pattern exists to close that gap: the system detects an unanswered call, sends a short message from the same number or a linked number, and logs the conversation in the CRM so someone can follow up. The idea is sound. Speed to the first reply is one of the few levers a small operations team controls directly.

The risk is treating the automation as a reason to skip the compliance steps. A text sent by software to a consumer's mobile number sits inside a body of federal regulation, and the penalties are statutory, applying per message. The Telephone Consumer Protection Act at 47 U.S.C. 227 is where the rules start, and the implementing regulation is 47 CFR 64.1200. This guide describes an operating process built around those texts and common carrier practice. It is not legal advice, and the rules, including how revocation of consent must be handled, have been amended in recent years, so read the current text and have counsel review the flow before launch.

One point deserves a flag because it is often assumed away. People sometimes argue that a caller who dialled your number has invited a reply. That may be a reasonable position for a single, immediate, informational response, but it is an interpretation, not a safe harbour that covers a sequence of marketing messages, a later promotion, or a follow-up days afterwards. Design the flow so that its first message is strictly responsive and carries no promotion. Treat any further messaging as needing clear consent, captured and stored.

Start with the consent record. For each phone number that receives automated messages, the business should be able to produce how consent was given, when, for what kind of messages, and from which page, form, or call. A web form checkbox that is not pre-ticked, with disclosure of the sender, the type and frequency of messages, that message and data rates may apply, and how to stop, is a common way to capture it, and a sensible baseline that fits the CTIA guidance on clear calls to action. Marketing texts typically need a stronger form of consent than purely informational ones under the FCC's rules, and the definitions turn on the wording of the rule, so check the current section before assuming a category.

Carrier and industry rules sit on top of the law. The CTIA's Messaging Principles and Best Practices describe expectations that carriers and aggregators enforce, including clear consent, identification of the sender, and honouring opt-outs through standard keywords such as STOP. Business messages sent over standard long codes may also need registration through the messaging provider (for application-to-person traffic in the United States, providers commonly require brand and campaign registration), and carriers may filter or block traffic they treat as unwanted, which is a risk for unregistered senders. Your provider's documentation, not this guide, is the authority on current registration steps.

Opt-out handling is where operations most often fail. When someone replies STOP, or any clear equivalent such as "unsubscribe" or "no more texts", the number must stop receiving messages from every system that can send to it, not just the one that received the reply. That means the text-back tool, the CRM sequences, the review-request flow, and any appointment reminder tool all need to share a suppression list or sync to the same one. Process opt-outs immediately. The FCC's revocation rules set outer limits, and they have changed, so read the current text, but there is no business benefit in waiting for the limit. Confirm the opt-out with one short message and then stay silent.

Keep the first message short, factual, and specific to the situation. It should identify the business, say that you missed the call, say what happens next, and offer one easy action. Avoid claims you cannot meet, such as "we will call in five minutes" outside staffed hours. If you operate with a callback window, state it. If the sender is an automated system, do not pretend otherwise. A caller who learns mid-conversation that it was a bot tends to lose trust, which is a judgement, not a measured finding.

After-hours scripts need a branch for emergencies. The simplest and safest design asks one question, and gives the caller a route that does not depend on the automation working: for emergencies that threaten life or safety, call the local emergency number; for urgent service matters, call the on-call line. A keyword reply such as URGENT can notify the on-call person and open a task in the CRM. Do not ask the caller to describe the emergency in a text and wait for a response. That puts a safety-critical decision on a slow channel.

Email follow-up for the same enquiry falls under CAN-SPAM when it is commercial. The FTC's compliance guide says commercial messages must not use false or misleading header information or deceptive subject lines, must identify themselves as ads where applicable, must include the sender's valid physical postal address, and must provide a clear way to opt out that the sender honours within ten business days. Transactional or relationship messages, such as a confirmation of a booking the customer made, are treated differently, but the false-header rule still applies to them. If a message mixes a confirmation with a promotion, the safest approach is to treat it as commercial.

Every missed call should create or update a record, whether or not a text goes out. The minimum fields are the caller number, the timestamp, the line dialled, whether the call was inside or outside staffed hours, the message sent, any reply received, consent status, and the owner assigned for follow-up. A call from an existing customer should attach to their record and not create a duplicate lead, so match on phone number before creating anything new. If the CRM cannot match, route to a human to check, and do not let the automation guess.

Assign a response owner and a response-time target, and write both down. Treat the target as an internal service level, not an industry standard: a common starting point is first human contact by the next business morning for after-hours enquiries and within the hour for business-hours misses, adjusted to your staffing. Surface any lead that has not been touched within the target on a morning list, since an untouched record is the main way text-back systems end up creating false confidence.

Log the compliance events as well. Consent source, opt-outs, and suppression syncs all belong in the record, because if a customer complains or a carrier asks questions, the business will need to reconstruct what was sent and why. Keep message content in the log as sent. Retention periods are a decision for counsel and the privacy policy, so apply the period you have documented rather than deleting by habit or keeping everything forever.

Decide in advance what counts as an emergency for your trade, and put the definition in writing: gas smell, active flooding, no heat in freezing conditions for a vulnerable occupant, an electrical fire hazard, or whatever applies. The text-back flow should never be the only thing standing between a caller and help. The first message after hours should include the emergency-number instruction, and the voicemail greeting should say the same thing. Test both with a real call every month, since forwarding rules, holiday schedules, and number porting quietly break them.

On-call routing needs a name, a phone number, a backup, and a rule for what happens if neither answers within a set number of minutes. Write the failure case down: if no one confirms an urgent request, the system sends the caller a second message with the emergency-number instruction and notifies a manager. Treat unanswered urgent requests as incidents, reviewed the next business day.

This process does not make a business reachable if no one is working the queue, and it does not resolve a consent problem created earlier, such as a purchased list or numbers collected without disclosure. It also does not replace a voice agent or an answering service where the volume justifies one. Those are covered under the voice agent launch guide. What it does is make the first reply fast, honest, and recorded.

Run the steps in order. Do not enable sending until the compliance gates have been shown to pass on a test number.