Chip Pro Autoglass replaced two systems with an AI receptionist that quotes and books windshield repairs

01

The Company

Chip Pro Autoglass is a mobile windshield repair company operating in Utah, run by Brad Haralson. The business is built on a specific position: repairing damage — including longer cracks — in cases where many shops would recommend a full replacement.

That position shapes everything about how Chip Pro sells. A repair has to be assessed before it can be priced, the customer often needs convincing that repair is even an option, and the whole exchange has to happen quickly enough that the lead is still warm. Response time to a new lead is the metric Brad manages the business by, and in auto glass the industry norm is slow enough that being fast is a genuine commercial advantage.

02

The Challenge

Chip Pro ran two platforms: Podium for customer messaging, and HouseCall Pro for field service management. Between them sat a manual seam. Paid social leads were routed through an automation tool into the field service inbox, and everything after that was a person carrying information between systems.

Windshield damage makes fast response harder than it sounds. It cannot be quoted sight unseen, and the lead form could not carry a photo. So every inbound enquiry meant a manual exchange to get pictures, a judgement call on whether the damage was repairable at all, and only then a price. A meaningful share of enquiries were never going to convert, and the team spent the same effort finding that out.

The core problems were immediate and critical:

• Two systems, one manual seam. Messaging lived in one platform and the job lived in another, and a person moved information between them on every enquiry.
• No photo on the lead form. Damage cannot be assessed from a text description, so every enquiry began with a manual round trip to get pictures.
• Effort spent on enquiries that could never convert. Whether damage was repairable at all was only knowable after the same work had already been done.
• A hard migration deadline. The Podium account was closing, and the entire customer history — contacts, conversations and photos — lived inside it.
• Quotes that could not be walked back. People routinely misjudge the size of the damage they are describing, so a precise price given early is almost impossible to correct without losing the job.

03

The Solution: An AI Receptionist That Quotes the Range and Books the Slot

Chip Pro replaced Podium and HouseCall Pro with FieldCamp, and put an AI receptionist in front of every inbound text channel. Brad configured most of it himself — writing the agent’s instructions, supplying a library of real damage photos, and defining where the agent should stop and hand over to a person.

What made FieldCamp different:

• An agent the customer configures, not a script the vendor writesBrad wrote the instructions, supplied the reference photo library, and set the conditions under which the agent stops and passes the conversation to a person.
• Quoting a range instead of a numberThe agent offers a bracket, explains that the technician confirms on arrival, and makes the booking risk-free — so the conversation reaches a booking without anyone committing to a price that cannot yet be stood behind.
• Real slots, offered in a drivable orderThe agent reads genuine open times from the team’s schedule and offers them route-optimised, establishing location before day and time.
• One inbox for every channelCalls, texts, email, online booking, the website widget and paid social leads arrive in the same place, with the team’s own tags and a clear owner on every conversation.
• A one-off migration bridge built for the deadlineFieldCamp wrote a direct import from Podium and ran it across a weekend before the account closed, with nothing dropped.

Moving Off Podium in a Single Weekend

Moving off Podium came with a hard deadline: the account was closing, and Chip Pro’s customer history lived inside it. FieldCamp built a one-off bridge into Podium, handled the authorisation with Brad, and ran a throttled, resumable import across the weekend. Every contact, conversation and customer photo came across before the cutoff, with no errors and nothing dropped.

An Agent That Quotes a Range, Not a Number

The FieldCamp AI receptionist works over text — SMS, paid social lead ads, and the website widget — and it quotes a range rather than a number. This was Chip Pro’s own design rule, and the reasoning behind it is the most transferable thing in the build.

People consistently misjudge the size of the damage they are describing. A crack reported as small routinely turns out to be several times that. A precise quote given early is almost impossible to walk back without losing the job. So the agent offers a bracket, explains that the technician confirms on arrival, and makes the booking risk-free: no obligation if the damage turns out to be outside what can be repaired.

Booking Real Times, in a Drivable Order

The agent books real times. It reads genuine open slots from the team’s schedule and offers them route-optimised, establishing location before day and time — because the reverse order produces a calendar that is expensive to drive. It also follows up on quiet leads by itself, on a cadence the business controls, and stops the moment the customer replies, books, opts out, or a person steps in.

Knowing When to Hand Over

Chip Pro sets the conditions under which the agent stops and passes the conversation to a human — including, deliberately, whenever the agent is not confident. Within weeks of going live, the request was not for tighter control. It was for the agent to respond automatically rather than waiting to be handed a conversation.

Call Copilot on the Human Side

Alongside the receptionist, FieldCamp Call Copilot works the human side of the business. The team answers live calls inside FieldCamp, and the AI extracts the caller’s details into the conversation record afterwards, so a call leaves the same trail as a text.

04

The Impact: From Two Systems to One Inbox

One Inbox Instead of Two Systems

Calls, texts, email, online booking, the website widget and paid social leads now arrive in one place. Conversations carry the team’s own tags, and each can be claimed, assigned to a person, assigned to the AI, snoozed to a date, or resolved.

The change that did most for response time is small and unglamorous: FieldCamp writes lead-form answers into the conversation as an internal note, so whoever picks it up can quote without opening another screen.

The FieldCamp website text widget does two things Chip Pro’s previous one did not — every submission creates a Request with the customer already attached, and the AI agent can send the first reply within seconds, while the lead is still warm. FieldCamp also texts back automatically on a missed call, so no caller goes cold.

Built With the Customer, Not For Them

Chip Pro sent months of near-daily feedback. What matters is what happened next. On one day in July 2026 they asked for seven inbox changes: mark as unread, inline contact editing, a send-and-close action, full-screen compose, a To-box that accepts names, multi-assign, and clearer labels. All seven were live within twelve hours.

“Hold this time for a lead” — greying out a slot while a customer decides — did not exist when Chip Pro asked for it. It shipped eleven days later, alongside arrival windows. Their one refinement, that the hold should not require an existing client record because sometimes people simply call, was live eight hours after they raised it.

Then there is the detail that says most about how the relationship works. Chip Pro mentioned their SMS bill looked high. FieldCamp went through the message library and found thousands of curly apostrophes — a single character that silently cuts SMS capacity by more than half and quietly inflates every bill.

The character set was normalised, automatic MMS switching was added above a threshold Chip Pro proposed, and a live segment counter went into the composer so the team can see the cost as they type. Nobody had asked for that. It came from reading the bill more carefully than the customer had.