Case study: home services
Answering every call, in every language.
A family-owned home-services company was losing jobs to missed and after-hours calls, and had no way to serve a fast-growing base of customers who did not speak English. AI went to work across the front office and the back office, and the business moved into growth mode. The client name is withheld to protect confidentiality.
Attribution, stated up front. This engagement was delivered under Redline Management, an AI advisory practice led by the same operator who leads Edgility Solutions. The same method and the same operator now run inside Edgility as the AI Implementation program. We are showing it here because it is the work, not because Edgility performed it as an entity.
The results
One engagement, across about two quarters.
Results specific to this client and this business. They are reported here as what happened, not as what a different business should expect.
Separately, and not a result of this engagement: industry sources including ServiceTitan, PHCC and Housecall Pro report that winning a customer through neighborhood referral costs roughly six times less than through paid advertising. That benchmark is context for why the referral and geofencing work below matters, and it is not a figure this client produced.
What was built
One system that answers, books, collects and defends margin.
Wired into the tools the team already ran on. The book of record did not move.
Answer
The phone stopped being a bottleneck
A multilingual voice agent answers every call, 24 hours a day, in the caller's language. It triages the problem, gives an estimate where one is appropriate, and books the job straight into the CRM rather than taking a message.
Collect
The money moved at the same moment
A deposit captured and the payment routed at the point of the booking, with a compliant pricing structure recovering card cost on every transaction, wired into the business management platform rather than sitting beside it.
Defend
Somebody finally checked the invoices
An agent checks every distributor invoice against live warehouse pricing at pickup and flags the difference for dispute. That work alone accounted for the 18 percent reduction in overbilling.
The growth flywheel
Every finished job markets the next one on the same street.
After the work is done, the same system turns one satisfied customer into a neighborhood of them. A satisfaction message routes happy customers toward a public review and unhappy ones straight to the owner, before they post anywhere.
The address of each completed job is geofenced so the neighbors around it see marketing while the truck is still memorable. The homeowner gets a shareable referral link that pays them when a neighbor books. Every job posts automatically to the social channels the business had never once used.
The effect: new jobs cluster by neighborhood, and the customer becomes the marketing.
Why it held
The part that is not software
A system like this fails when the dispatcher keeps answering the phone the old way and the office manager never opens the invoice queue. Adoption was coached rather than assumed, and the way the day runs changed with it.
The money side is the other half. Deposits, routing, settlement, reconciliation and the pricing structure are payments work, not software work, and they are where a general AI build tends to stop at the demo.
Guardrails were set at design time: scoped access, defined escalation to a human, and the pricing structure disclosed transparently and in line with card network rules.
Start here
This could be your business.
If the business runs on inbound calls, the same leak is costing you jobs and margin right now. Twenty minutes to map it, at no cost.