Warranty and Callback Handling

A callback and a new inquiry look identical when the phone rings. Someone says the heating is not working, the office books it, a technician attends, and only at the property does anyone realize this is the same system the business installed eleven weeks ago. By then the visit has been scheduled as revenue, the technician has been pulled off paid work, and the awkward conversation about who pays happens on the customer's doorstep with no record of what the warranty actually covers.

Because callbacks are not separated at the point of contact, three things stay permanently unknown. The business does not know how many of its visits are callbacks, so the true cost of rework is invisible and gets absorbed into general labor. It does not know why jobs come back, because nobody records a cause code, which means the same installation defect recurs across dozens of jobs without anyone noticing a pattern. And it does not know which technicians, which trades or which product lines generate the rework, so coaching and supplier decisions are made on impressions.

The money leaks in both directions, which is what makes this worth building properly. Genuine warranty work gets billed to customers who then dispute it, which costs a review and sometimes the relationship. Chargeable work gets written off as goodwill because nobody could prove what was in scope, which is a direct hit to margin repeated quietly every week. Manufacturer warranty claims go unfiled because the window closed before anyone gathered the serial number and the fault description, so the business eats a component cost the manufacturer would have covered. And repeat callbacks - the second and third visit to the same job - are the most expensive visits a home services business makes and almost never appear on a report, because every one of them is logged as a fresh job.

Pipeline and stages

A separate pipeline, Warranty and Callbacks, kept deliberately apart from the sales pipeline. A callback is not a lead and must never inflate pipeline value or conversion reporting.

• Callback Reported - entry stage, created at first contact.
• Triage - Classifying - original job located, warranty dates being checked.
• Warranty Confirmed - determination is Covered; no charge conversation needed.
• Chargeable - Awaiting Approval - determination is Not covered; customer must approve before a visit is booked.
• Visit Scheduled - appointment exists with a technician attached.
• On Site - technician has arrived; the on-site determination form is pending.
• Resolved First Visit - fixed in one attendance. This is the stage whose share you are trying to grow.
• Escalated - Repeat Callback - second or later visit on the same original job. Cards only enter here automatically.
• Manufacturer Claim Open - a claim has been filed and a reference recorded.
• Closed - Warranty, Closed - Chargeable, Closed - Goodwill - three separate closed stages, because collapsing them into one destroys the only report that matters here.

Calendars

• Callback Visit: a service calendar with two slots reserved every weekday morning, released to normal booking at 16:00 the previous day if unused. Reserving capacity is what stops callbacks either waiting a week or cannibalising booked revenue at random.
• Warranty Specialist: a round-robin across the senior technicians who hold authority to make an on-site determination. Repeat callbacks route here rather than back to the original technician, because the person who missed it the first time is the wrong person to diagnose it the second.
• Manufacturer Attendance: a collective calendar requiring both a company technician and a manufacturer engineer, used when a claim needs joint attendance.
• First-callback visits are weighted to the original technician where availability allows, which is both fair and instructive - a technician who returns to their own work learns faster than one who is shielded from it.

Custom fields

• Original Job Reference - text. The spine of the whole build. Nothing can leave Triage without it.
• Original Job Date and Original Technician.
• Trade Category - dropdown: HVAC, Plumbing, Electrical, Roofing, Appliance, Other.
• Warranty Type - dropdown: Labor warranty, Parts warranty, Manufacturer warranty, Out of warranty, Unknown.
• Labor Warranty Expiry, Parts Warranty Expiry, Manufacturer Warranty Expiry - three separate dates, because they almost never match and a single warranty date field is the most common reason a business bills the wrong party.
• Callback Reason Code - dropdown: Installation defect, Component failure, Incomplete work, Commissioning not completed, Diagnostic error, Customer misuse, Unrelated new fault. Mandatory at close. This single field is the reason to build the system.
• Callback Sequence Number - number. 1 for a first return, 2 for the next on the same Original Job Reference, and so on. Calculated automatically.
• Warranty Determination - dropdown: Covered, Not covered, Goodwill, Manufacturer claim.
• Determination Made By - user reference - and Determination Date. Who decided and when, every time, without exception.
• Chargeable Approval Received - switch, with Approval Method alongside it.
• Cost Owner - dropdown: Company, Manufacturer, Customer, Subcontractor. Who bears it, not what it costs.
• Manufacturer Claim Reference and Serial or Model Number.
• Days From Original Job - number, calculated at creation. The number that drives nearly every branch in the automations.
• Resolution Notes - long text, written on site, not from memory the next morning.

Tags

• callback-reported - applied at creation, so callbacks are excluded from marketing sends for the duration.
• callback-warranty, callback-chargeable, callback-goodwill - mutually exclusive, applied at determination.
• callback-repeat - applied automatically when Callback Sequence Number reaches 2.
• callback-escalated - applied at sequence 3, which is the point a manager gets involved personally.
• manufacturer-claim-open
• resolved-first-visit
• suppress-review-request - applied at creation and removed only on a clean resolution. Asking for a review while a callback is open is the fastest way to generate a bad one.
• warranty-expiring-60 - applied by a scheduled scan, used for the proactive pre-expiry check rather than any callback itself.

Forms

• Callback Report - used by whoever takes the call, or sent to the customer as a link. Fields: customer lookup, original job reference, what is happening now, when it started, whether anything has changed at the property since the work, whether there is a safety issue, photo or video upload, and best contact window. The has anything changed question resolves a surprising number of misuse cases before a van leaves the yard.
• On-Site Callback Determination - technician-facing, mandatory before the card can close. Fields: Callback Reason Code, Warranty Determination, Cost Owner, serial or model number, what was actually done, photographs of the fault and of the repair, and whether a part is needed. If a part is needed the submission hands straight over to the parts hold process rather than inventing a second tracking method.
• Repeat Callback Escalation - triggered at sequence 2 and above. Captures what the previous visit concluded, why that was wrong, and what is being done differently. Filled by the senior technician, reviewed by the operations owner weekly.

BO's role and guardrails

BO does triage, not adjudication. On an inbound message BO identifies whether this is new work or a return, finds the original job, and reads the three expiry dates.

• BO may ask for the original job reference or approximate date, identify the matching job, confirm receipt, and state factually whether the recorded warranty dates have passed.
• BO may never make or imply a determination. BO does not say that'll be covered or that sounds like wear and tear. Determination belongs to a named human, on site, recorded in Warranty Determination with their name against it.
• BO may never discuss charges, call-out fees, or what the customer will or will not pay.
• Immediate human handoff on: any safety indicator - no heat, no water, no power, gas smell, water escaping, electrical burning smell - which bypasses BO entirely and rings the on-call group; any mention of a lawyer, regulator, trading standards, insurer or review site; any second or later callback; any abusive or highly distressed message.
• BO always states that a human will confirm the warranty position, and never leaves the customer believing the matter is settled.

Numbers and channels

• The main business number carries an IVR option - press 2 if you are calling about work we have already carried out - which routes to a dedicated ring group and tags the call before anyone speaks. This one change does more for callback data quality than any report.
• SMS on the main number for customer updates and the chargeable approval link, threaded into the shared inbox against the contact so whoever picks it up sees the original job history.
• A dedicated warranty email address for manufacturer claim correspondence, parsed onto the opportunity so claim threads live with the job rather than in one person's mailbox.
• A #callbacks internal channel that posts only three things: a new repeat callback, a safety-flagged callback, and the Monday digest. Nothing routine.

1. The customer calls and the IVR sorts them before anyone speaks. They press the option for work already carried out. The call routes to the callback ring group rather than the general queue, and the call is tagged. If they message instead, BO does the same job in text.
2. The original job is found, not guessed. Whoever answers - or BO, on a message - locates the original job and the Callback Report form is completed from the conversation. The form captures what is happening now, when it started, whether anything has changed at the property, and whether there is any safety concern. A photo or video is requested while the customer is still on the line, because they will not send one later.
3. Safety overrides the whole process. If a safety indicator was flagged, everything in this list collapses: the on-call group is rung immediately, nothing waits, and the determination conversation happens after the property is safe. This path is built first and tested first.
4. A card is created and the customer is acknowledged within minutes. An opportunity appears at Callback Reported, the three warranty expiry dates are pulled from the original job, Days From Original Job and Callback Sequence Number are calculated, and the customer gets a text confirming it has been logged against a specific previous job on a specific date. Naming the original date is what tells the customer they are known.
5. A named human makes the determination. The card moves to Triage - Classifying and the office owner reviews the three expiry dates against the reported fault. They set Warranty Determination to Covered, Not covered, Goodwill or Manufacturer claim. The system stamps their name and the date. No card advances on an anonymous determination.
6. Covered work goes straight to a booking link. The customer gets a text saying it is covered, there is nothing to pay, and here are the available slots - weighted to the original technician, drawn from the reserved morning capacity on the Callback Visit calendar.
7. Chargeable work gets a phone call before it gets a link. If the determination is Not covered, nobody sends a booking link. The office calls, explains which warranty had lapsed and when, and only then sends the written confirmation and the approval link. The written record of that conversation is what prevents the dispute.
8. The visit is booked and reserved capacity absorbs it. Because two morning slots are held back each weekday, the callback does not displace booked revenue or wait a week. Unused slots release to normal booking at 16:00 the day before.
9. The technician attends and closes out on site. Thirty minutes after the appointment ends, the technician gets a link to the On-Site Callback Determination form. They record the reason code, confirm or correct the determination, name the cost owner, capture the serial number, photograph the fault and the repair, and write what they actually did. This is the step that produces every number in the reporting, so it is chased until it is done.
10. A contradiction routes to a manager, not to a close. If the technician's on-site determination differs from the office determination, the card goes to the manager. The office saw a date; the technician saw the fault. The technician is usually right, and the record should show the change and who made it.
11. A part requirement hands over cleanly. If the fix needs a component, the submission creates a parts hold using the existing parts process rather than a second, parallel tracker. The callback card waits at its stage and resumes when the part is fitted.
12. First-time fixes close; repeats escalate. Resolved at sequence 1 closes to the matching closed stage with a reason code recorded. Unresolved increments the sequence number, which immediately routes the card to a senior technician, excludes the original technician, posts to the internal channel, and tasks the operations owner with a personal phone call.
13. The review request is gated on a direct question. Three days after close the customer is asked one yes-or-no question about whether things are still working. Only a yes releases the review request, and only for clean first-visit resolutions. A no reopens a callback card at the next sequence number.
14. Monday morning the pattern gets looked at. A single digest lands in the internal channel: callbacks opened and closed, first-visit fix rate, reason code distribution, repeats by trade and technician, open manufacturer claims. The reason codes are the agenda for the week, which is the entire point of having collected them.

1. Callback Intake and Triage

Trigger: Callback Report form submitted, or an inbound call tagged by the IVR existing work option, or BO classifying a message as relating to previous work.

Branches: first on the safety question. Any yes - no heat, no water, no power, gas smell, water escaping, burning smell - skips every wait in this workflow, rings the on-call group immediately, and posts to #callbacks within seconds. Second branch on whether Original Job Reference could be matched automatically. Third on Days From Original Job: under 7 days routes as an incomplete-work follow-up rather than a warranty question, because a job that fails inside a week is almost always commissioning.

Actions:

1. Create an opportunity in Warranty and Callbacks at Callback Reported. Apply callback-reported and suppress-review-request.
2. Match the original job, populate Original Job Reference, Original Job Date, Original Technician, Trade Category and the three expiry dates. Calculate Days From Original Job and Callback Sequence Number by counting existing cards against the same reference.
3. Wait 3 minutes. SMS the customer: "Thanks [first name] - we've logged this against the work we did for you on [original job date]. Someone from our team will confirm the next step with you within [X] business hours. If anything feels unsafe, call us straight away on [number]."
4. Move the card to Triage - Classifying and create a triage task for the office owner due in 2 business hours.
5. If no original job could be matched, the task instead reads Identify original job before triage and the card is held. A callback with no original job is an unmeasurable callback.
6. If Callback Sequence Number is 2 or more, skip the rest and hand straight to workflow 5.

Exit: card reaches a determination stage, or is reclassified as new work - in which case it closes here and an opportunity is created in the sales pipeline instead, with a note explaining the reclassification.

2. Warranty Determination Routing

Trigger: Warranty Determination is set on a card in Triage - Classifying.

Branches: four ways on the determination value. Covered goes straight to booking. Not covered goes to the approval workflow. Goodwill books like Covered but requires a manager to be the Determination Made By user, which is enforced rather than requested. Manufacturer claim forks into workflow 6 and books a visit in parallel.

Actions:

1. Stamp Determination Date and Determination Made By. A determination with no name attached is rejected and the card stays in triage.
2. Apply the matching tag - callback-warranty, callback-chargeable or callback-goodwill - and remove any previously applied determination tag so the three stay mutually exclusive.
3. Set Cost Owner from the determination, defaulting to Company for Covered, Customer for Not covered, Manufacturer for a claim.
4. Covered path: move to Warranty Confirmed and SMS "Good news - this is covered under the warranty on your original job, so there's nothing to pay. Pick a time that suits here: [booking link]."
5. Not covered path: move to Chargeable - Awaiting Approval and hand to workflow 3. No booking link is sent yet. Sending a booking link before the charge conversation is how doorstep disputes start.
6. Write a note recording which of the three expiry dates drove the decision, so the reasoning survives staff turnover.

Exit: card reaches Visit Scheduled, or the determination is reversed - which reverses the tags and reruns this workflow from the top.

3. Chargeable Approval and Booking

Trigger: stage changes to Chargeable - Awaiting Approval.

Branches: on Days From Original Job. Inside 30 days past the relevant expiry, the first message is softer and a goodwill review task is created for the manager before any chase runs, because a job that failed days after cover lapsed is a relationship decision, not a billing one. Beyond that, the standard ladder runs.

Actions:

1. Create the chargeable approval task for the office owner, due in 2 business hours: call the customer first, do not lead with a text.
2. After the call is logged, send the written confirmation by SMS and email with the approval link: "As discussed, [first name] - the fault we'd be attending isn't covered by the warranty on your original work, so this visit would be chargeable. Approve here and we'll get you booked: [link]. Happy to talk it through again if that's easier."
3. Wait 1 business day. If Chargeable Approval Received is on, move to Visit Scheduled and send the booking link. Exit.
4. Wait 2 business days. Second SMS: "Still here if you want to go ahead with that visit - approve any time here: [link]."
5. Wait 3 business days. Create a final call task for the manager, framed as a decision: proceed chargeable, offer goodwill, or close.
6. Wait 5 business days. Close as Closed - Chargeable with a declined note, and remove suppress-review-request only if the final call was logged as amicable.

Exit: approval received, manager converts the card to Goodwill, or the ladder completes.

4. On-Site Determination Capture and Close-Out

Trigger: appointment on the Callback Visit or Warranty Specialist calendar moves to showed, which sets the card to On Site.

Branches: on the on-site form answers. If the technician flags that a part is needed, the card hands over to the parts hold process and does not close. If the on-site determination contradicts the office determination, the card routes to the manager rather than closing on the technician's word alone.

Actions:

1. Wait 30 minutes after the appointment end time, then SMS the technician a link to the On-Site Callback Determination form: "Callback close-out for [customer] - two minutes while it's fresh: [link]."
2. Wait 2 hours. If unsubmitted, resend. Wait until 18:00, then create a task for the technician's supervisor. The form is the entire dataset; chasing it is not optional.
3. On submission, write Callback Reason Code, Cost Owner, serial or model number and Resolution Notes to the opportunity, and attach the fault and repair photographs.
4. If resolved and Callback Sequence Number is 1, move to Resolved First Visit, apply resolved-first-visit, then close to Closed - Warranty, Closed - Chargeable or Closed - Goodwill matching the determination.
5. If not resolved, increment Callback Sequence Number and hand to workflow 5.
6. Wait 1 hour after close, then SMS the customer: "All done - [technician] has written up what was found and what was fixed. If anything isn't right, reply to this message and it comes straight to us."

Exit: card closed, or diverted to parts or escalation.

5. Repeat Callback Escalation

Trigger: Callback Sequence Number reaches 2 or more on any card.

Branches: sequence 2 escalates to the operations owner. Sequence 3 or more escalates to the business owner and the manufacturer claim path is reviewed regardless of the original determination, because three visits usually means a component problem rather than a workmanship one.

Actions:

1. Move the card to Escalated - Repeat Callback. Apply callback-repeat; at sequence 3 also apply callback-escalated.
2. Post to #callbacks immediately: customer, original job, trade, sequence number, previous reason codes. Repeats are one of only three things allowed to post to that channel.
3. Route booking to the Warranty Specialist round-robin and exclude the original technician from assignment.
4. Create a task for the operations owner due in 4 business hours: Call customer personally before the visit is booked. A second failure is a relationship event and a text will not hold it.
5. Send the Repeat Callback Escalation form to the assigned senior technician, required before the visit.
6. Wait until the visit is completed. Create a review task for the operations owner within 2 business days: confirm the reason code, confirm the cost owner, and decide whether this pattern needs a toolbox talk, a supplier conversation or a product line review.

Exit: resolved and closed with a reason code recorded. Repeat cards never auto-close; a human closes every one.

6. Manufacturer Claim Tracking

Trigger: Warranty Determination set to Manufacturer claim, or Cost Owner set to Manufacturer.

Branches: on whether Serial or Model Number is populated. Without it the claim cannot be filed, so the workflow blocks and tasks the technician rather than filing an incomplete claim that gets rejected weeks later.

Actions:

1. Move to Manufacturer Claim Open; apply manufacturer-claim-open.
2. Create a task for the warranty administrator due in 1 business day: File claim, with serial number, fault description, photographs and original job date assembled in the task body so nothing is hunted for.
3. On Manufacturer Claim Reference being populated, email the customer a short confirmation that a claim is open with the manufacturer and that the repair is not waiting on it.
4. Wait 7 days. If no manufacturer response is logged, create a chase task and email the manufacturer from the warranty address quoting the reference.
5. Wait 14 days. Second chase, and post to #callbacks.
6. Wait 30 days. Escalate to the operations owner to decide whether to pursue, absorb, or raise with the supplier relationship manager. Record the outcome in Cost Owner, so absorbed claims are countable.

Exit: claim settled or formally abandoned with the outcome written to Cost Owner.

7. Review Request Gate and Resolution Follow-Up

Trigger: card moves to any closed stage.

Branches: on closed stage and sequence number. Closed - Warranty or Closed - Goodwill at sequence 1 is eligible for a review request. Anything at sequence 2 or more, and anything closed as chargeable after a declined approval, is permanently excluded - the suppress-review-request tag is simply never removed.

Actions:

1. Wait 3 days after close. Send a single satisfaction question by SMS: "Quick one, [first name] - is everything still working as it should after our visit? Just reply yes or no."
2. A no reply creates an urgent task, reopens a new callback card at the next sequence number, and keeps the suppression tag in place. A no is the cheapest early warning available.
3. A yes reply removes suppress-review-request and callback-reported, then waits 1 day before sending the review request so the two messages are not stacked.
4. No reply after 4 days removes the suppression tag but sends no review request. Silence is not consent to be asked.
5. Write the final reason code and cost owner to the contact record so the next person who opens it can see the history without reading the timeline.

Exit: review request sent, or permanently suppressed.

8. Pre-Expiry Warranty Check

Trigger: weekly on Monday at 09:00, scanning completed jobs whose Labor Warranty Expiry falls 60 days out.

Branches: on Trade Category, because what is worth proactively checking differs by trade and the message should name something specific rather than being generic. Exclude any contact with an open callback card or carrying callback-repeat.

Actions:

1. Apply warranty-expiring-60.
2. Send one email: "[First name], the labor warranty on the work we did for you on [date] runs until [expiry date]. If anything has been playing up, now is the time to tell us - while it is still covered. Reply to this email or book a check here: [link]."
3. Wait 7 days. If no reply and no booking, send one SMS with the same offer, then stop. One email and one text, then silence - this is a courtesy, not a campaign.
4. Any reply or booking creates a callback card at sequence 1 through workflow 1, so proactive finds land in the same dataset as reactive ones and the reason codes stay comparable.

Exit: expiry date passes, a callback card is created, or the contact opts out.

What is different afterwards

The business can finally answer the question it could never answer: what proportion of our visits are rework, and why. Because every callback carries a reason code set on site by the person who saw the fault, the causes separate out - installation defects from component failures from jobs that were never properly commissioned from faults that were nothing to do with the original work at all. Those four things need four different responses, and previously they were one undifferentiated pile of annoyance.

Billing stops being a doorstep argument. Determination happens before the van moves, by a named person, against three distinct expiry dates, with the reasoning written down. Chargeable visits are approved in writing before they are booked. The dispute rate falls not because the business bills differently but because the customer is told the position while they are still calm and before anyone has been to the house.

Repeat callbacks become visible and expensive-looking, which is what they are. A second visit on the same job now routes away from the technician who went the first time, triggers a personal call from a manager, and appears in a digest. That changes behavior fast, in both directions - technicians get better, and genuinely faulty product lines get escalated to suppliers with evidence attached.

Reputation stops being collateral damage. Because review requests are suppressed while a callback is open and only released on an explicit confirmation that things are working, the business stops soliciting reviews from precisely the customers least inclined to write a good one.

Metrics to watch

• Callback rate - callbacks as a share of completed jobs, tracked by trade category and by original technician. Measure the rate, not the count, or growth will look like decline.
• Reason code distribution, monthly. This is the single most actionable report the build produces.
• First-visit fix rate - share of callbacks closing from Resolved First Visit. The headline operational number.
• Repeat callback rate - share of callbacks reaching sequence 2 or higher, and the average sequence number per original job.
• Days From Original Job distribution. A cluster inside the first two weeks points at commissioning; a cluster at the eleven-month mark points at something else entirely.
• Cost owner split - Company versus Manufacturer versus Customer versus Subcontractor, by trade. Watch the Company share.
• Goodwill frequency, and who authorized it. Goodwill is a legitimate tool and a terrible habit; counting it by authorizer keeps it honest.
• Manufacturer claim recovery rate - claims filed versus claims settled, and average days to settle by manufacturer.
• Determination reversal rate - how often the on-site determination contradicts the office determination. A high rate means the triage step needs better information, not that technicians are difficult.
• Time from report to resolution, split by determination, and time from report to first customer contact, which is the number customers actually feel.
• Review rating on callback-affected jobs versus clean jobs, which tells you whether the gating is working.

Prerequisites

• Written warranty terms you can actually build to. Most businesses discover during this build that labor and parts cover have never been documented consistently, and that different technicians quote different periods. Settle it on paper first; the three expiry date fields will expose any ambiguity immediately.
• Historical job data with a usable job reference, completion date, technician and trade. Without a matchable original job the whole structure has no spine, so plan a data clean-up as part of the scope rather than a surprise during it.
• A named determination authority - who may set Covered, who may authorize Goodwill - and a deputy for each. The build enforces this, so it must be decided before, not during.
• An agreed reason code list that the technicians helped write. Codes imposed from the office get ignored or set to whatever is first in the dropdown.
• The list of senior technicians who can hold a determination, which populates the Warranty Specialist round-robin.
• Agreement on reserved callback capacity. Two morning slots a day is the suggested default; whatever is chosen has to be protected, or the calendar fills with new work and callbacks drift again.

Integrations

• Google Calendar or Microsoft 365 two-way sync for every technician, so reserved callback slots respect real availability.
• Google Business Profile and the other review destinations, connected for the gated review requests and for monitoring what gets written after a callback.
• Slack or Microsoft Teams for the #callbacks channel. Hold the line on posting only repeats, safety flags and the Monday digest.
• A dedicated sending domain with SPF, DKIM and DMARC in place. Manufacturer claim correspondence and written determinations both need to arrive and be findable later.
• QuickBooks or Xero where chargeable callbacks are invoiced, so Cost Owner drives whether an invoice is raised at all.
• Field service or job management software for the original job history. Where there is no native connector, pull a nightly export of completed jobs into the platform keyed on the job reference, and push callback outcomes back by outbound webhook. Agree the key before any build work starts.
• Manufacturer claim portals are rarely integrable. Treat them as a task-driven manual step with the evidence pre-assembled, which is most of the value anyway.

Number provisioning and messaging compliance

• Provision or port the main business number and build the IVR before anything else, since the press 2 for existing work option is what makes the data usable.
• A2P 10DLC registration is required for application-to-person SMS in the United States. Register the brand and campaign early and submit the real message samples, including the acknowledgement text, the chargeable approval text and the satisfaction check. Review timelines are outside your control.
• Decide explicitly which of these messages are transactional and which are marketing. The acknowledgement, the determination notice and the appointment reminders are operational. The pre-expiry warranty check and the review request are not, and must respect marketing opt-out status separately. Write that decision down and configure the workflows to match it.
• Honor STOP, STOPALL, UNSUBSCRIBE, END, QUIT and CANCEL automatically on every sequence, and make sure an opt-out never silences a safety-related message.
• Quiet hours on every timed workflow - roughly 8am to 8pm local - with only the safety escalation path permitted to override them.
• Identify the business clearly in the first message of every sequence, and keep approval and booking links on your own domain.

Data protection and record keeping

• Determination records - who decided, when, on what basis - are the evidence behind every billing decision in this build. Decide the retention period deliberately, and set it longer than the longest warranty you offer, not shorter.
• On-site photographs of faults and repairs are both operational evidence and images of the inside of someone's home. Restrict who can view them, set a retention period, and brief technicians to frame tightly on the fault.
• Call recording on the callback ring group is extremely useful for resolving disputes about what was agreed, and in some jurisdictions requires notice or the consent of all parties. Configure the announcement before enabling it.
• Manufacturer claims involve sharing customer and property details with a third party. Confirm what may be shared and under what basis.
• Restrict access to warranty determination and cost owner fields by role; not every user who can open a job needs to see what the business absorbed.

This is not legal advice. Warranty obligations, consumer protection rules, recording consent and messaging regulations vary by state and by country, and statutory rights frequently sit on top of whatever a business offers contractually. Have the client's own counsel confirm the warranty terms, the determination wording used in customer messages, recording notices and retention periods before go-live. To scope the build or the compliance review, book a call at bookings.appbo.io.