The cheapest revenue in a service business sits in quotes already sent and customers already served. Second contact rarely happens, because the phone rings and the day fills. AA19 builds the cadence so the pipeline stops depending on someone remembering.
AI improves follow-up by running a persistent cadence on unsold estimates, dormant customers, and stalled deals, written in the owner's voice and paused the moment a reply arrives. Revenue already earned stops leaking.
Follow-up loses to whatever is urgent. A quote goes out Tuesday, the intention to check in on Friday is real, and then a crew calls in sick. Unsold estimates pile up in a CRM nobody opens, customers pass their service interval unnoticed, and leads paid for through advertising go cold after one touch.
The expense already happened. Marketing generated the lead, an estimator drove to the property, and the proposal took an hour to build. Walking away after a single attempt writes off all of it, and the customer usually buys from somebody, just not from the company that quoted first.
A company sending forty estimates a month at a 35 percent close rate leaves twenty-six unsold quotes on the table each month. Three hundred a year, in a system nobody opens.
Add dormant customers past their service interval and dead leads from paid campaigns. The list is large, warm, and already paid for.
Messaging gets written in the owner's voice with real substance: a photo from a comparable completed job, a financing reminder, a seasonal price note, an availability opening next week.
Timing varies by ticket size and trade. A 900 dollar drain job follows a different rhythm than a 40,000 dollar roof with an insurance claim behind it.
A reply, a booking, or a card swipe pauses the sequence and notifies the right person. Nobody gets chased after saying yes.
Response data feeds back into the system. Sequences that convert on quotes above 10,000 dollars differ from the ones that recover a lapsed cleaning client, and the record shows which is which.
Corrections from the owner get preserved as rules, so tone drift and off-brand phrasing stop repeating.
Companies with a quote-to-close gap or a service interval. Roofers with unsigned proposals, remodelers with long decision cycles, HVAC dealers holding replacement quotes, med spas rebooking treatment series, dentists recalling overdue hygiene patients, insurance agencies working renewals, real estate agents nurturing past clients, and law firms following up on consultations.
First on the street after a storm
Scope written the same day
Maintenance base stays full
More bids submitted
Consultations that book themselves
Consultations booked from every inquiry
Fuller hygiene schedule
Quotes that close
Leads answered in minutes
Signed cases from the same lead volume
Dollars closed from quotes older than seven days.
Which message in the sequence earns the response.
Dormant customers returning to a booked job.
Deals sitting past the expected decision window, with escalation triggers.
| Area | Manual approach | AA19 system |
|---|---|---|
| Coverage of unsold quotes | Whichever ones someone remembers | Every open quote on a cadence matched to ticket size |
| Message quality | Varies by person and mood | Written in the owner's voice with job photos and real substance |
| Timing | Whenever the day allows | Sequenced by trade, ticket, and prior response behavior |
| Stopping the sequence | Customers chased after buying | Reply, booking, or payment pauses it and alerts the assigned person |
| Dormant customers | Contacted during slow weeks, sometimes | Segmented by last service date and worked continuously |
| Learning | Nothing recorded | Reply data by touch and segment feeding the next sequence |
Scope drives the number. Systems, trades, and custom pieces are all listed on the pricing page.
Messaging gets written from the owner's own language and job history. Generic drip copy does not go out.
Duration matches the ticket and the trade, from a five-day cadence on small service work to a multi-month rhythm on large projects.
Any reply, booking, or payment pauses the sequence immediately and alerts the assigned person.
Channels mix by customer preference and response history, with call tasks routed to a person.
Historic records get cleaned, deduplicated, and segmented by last service date, ticket size, and service line before anything sends.
One-time cost runs 500 to 10,000 dollars based on scope. A single unsold estimate cadence sits at the low end. Multi-segment work across quotes, dormant customers, reactivation campaigns, and referral requests sits higher.
Sequences run against the records already stored in ServiceTitan, Jobber, Housecall Pro, HubSpot, or a spreadsheet export. Replies, bookings, and payments write back so pipeline stages stay accurate without manual updates.
Copy gets drafted from the owner's own language and reviewed before a sequence activates. Corrections get preserved as rules, and reply data by segment shows which messages earn answers and which get retired.
Insurance claims, disputed invoices, warranty complaints, and accounts flagged by the owner get excluded from automated cadences and routed to a person. Any unexpected reply sentiment pauses the sequence and escalates.
Two to four weeks, with historic list cleanup usually driving the schedule. Records get deduplicated and segmented by last service date, ticket size, and service line before the first message sends.
All eight outcomes: More estimates, More booked jobs, Fewer missed calls, Less admin, Better follow-up, More reviews, Happier customers, Less time in the office.
Last updated: