Repair shop · Suwanee, GA
Norman's Electronics
A custom CRM, phone system and AI automation that cut the shop's monthly software and phone costs by about 80% and took hours of manual work off the staff every week.
About 80%
lower monthly software and phone costs
$6,700 to $8,000
saved per year
$392 → about $45
monthly phone bill
~2.5 hours a day
of "is it worth fixing?" calls now handled online first
The business
Norman's Electronics is an audio and appliance repair shop doing both customer-paid and manufacturer warranty repairs, on thin margins. Customers reach the shop by phone, text and email, often all three for the same repair.
The challenge
The shop was paying $836 a month for a CRM and phone system that didn't work together. The CRM was an old desktop database moved to the cloud, so every desk needed two monitors, and texting required a separate desktop app plus copying and pasting phone numbers to find the customer. The phone plan charged $392 a month, per seat and per number, and required a desk phone at every workstation. Customers called repeatedly for status updates, and hundreds of warranty invoices and credit memos arrived by email every week, sorted by hand. Salesforce quoted $10,000 in customization on top of monthly per-seat fees, and industry-specific software either ignored the shop's requirements or was built on a database design that would have created a maintenance problem.
Our approach
We sat at the staff's desks to learn how the work really happened, then fixed one bottleneck at a time. We started small with secure card-on-file payments, built everything on clean customer and repair data, and ran every AI step by hand until it was trusted. The first piece, secure card-on-file payments, was up and running for about $60 in AI build credits.
The cost difference
Occasional AI credit top-ups (about $60 each) apply during new feature builds or heavy use. Even with three in a month, costs stay about two-thirds lower than before.
What we built
The CRM
Every customer, repair, estimate and conversation lives on one screen. Customers approve estimates with one click, which triggers payment and the next step. Reports show revenue by service line and brand to guide pricing.
Result: One screen per desk instead of two, and every conversation in one place, so no one hunts across systems for a customer's history.
Texting integration
Status texts go out automatically when a repair moves forward, and customer replies land on the repair record and alert the right person. Status calls dropped the week it launched.
Result: Texting costs dropped by more than 70% per message, and status calls fell the week it launched. Customers can also text 411 to get the shop's hours for the week automatically.
Phone system
Calls ring in the browser from anywhere, and the caller's repair opens on screen automatically. One number handles unlimited calls at under a penny a minute, with no desk phones to buy or maintain.
Result: The phone bill went from $392 a month to about $45.
Voicemail triage with AI
Voicemails are transcribed and sorted by AI, robocalls are blocked before they reach the voicemail box, and staff can answer a voicemail with a text. Automatic text replies switch on once the staff trusts the AI's answers.
Result: Robocall voicemails stopped, and callers get answered faster without a phone call back.
Email triage and billing updates with AI
AI reads the billing inbox, finds ticket and PO numbers in the emails and PDF attachments, attaches each invoice to the right repair, and enters the parts costs automatically. It also matches warranty credit memos to their invoices and files emailed or scanned receipts and damage photos on the right ticket. Anything uncertain is flagged for a person to review.
Result: Hundreds of emails a week that were sorted by hand now file themselves, saving hours every week.
Customer website
Customers request service, pay deposits securely and look up status online. A service catalog built from past repairs shows which brands and models the shop works on and average repair costs, updated nightly. An AI advisor answers "is this worth fixing?" using the shop's real repair history and comparable used prices. It's kept separate from the CRM to protect customer data.
Result: About 8 twenty-minute calls a day now start online, and publishing average costs up front cut down on price-check calls.
Tools used
What's your bottleneck?
Thirty minutes, no pitch deck. Bring the thing that eats your week.