This hub uses collision-industry vocabulary on purpose. It's how shop owners talk. But if you're a controller, a partner, a lender, or anyone else evaluating the fleet, this glossary covers every term we leaned on, alphabetically.
The carrier's pre-loss valuation of a vehicle. Combined with the repair estimate, ACV determines whether a job is repaired or declared a total loss. If the estimate approaches the state's total-loss threshold as a percentage of ACV, the repair order ends and a settlement begins.
The sensor and camera systems — lane-keep, adaptive cruise, blind-spot, automatic emergency braking — that are standard on roughly 60% of 2025 model-year vehicles. ADAS calibration after collision repair is a specialized operation often performed as a sublet.
Enterprise's portal for coordinating loaner/rental authorizations between shops, carriers, and rental branches. Extensions require manual entry of delay reason codes and new return dates — which is exactly what Agent 03 automates.
One of the three major estimating platforms used in North American collision. Audatex (owned by Solera) exposes its integration surface through the OAS3 standard.
A fully electric vehicle with no internal combustion engine — Tesla, Rivian, Ford Lightning, Chevrolet Equinox EV, and so on. Collision work on BEVs requires high-voltage safety procedures that Agent 05 documents and bills automatically.
A paint operation where color is extended into an adjacent undamaged panel to avoid a visible transition line. Blend time is a P-Page item that is often under-written on estimates, making it a common target for Agent 01's scrub.
CCC Intelligent Solutions' estimating and workflow platform — the dominant system in North American collision, used by roughly 28,000 shops. CCC ONE is the source of most workfile events the fleet listens to.
CCC's developer integration surface (cccsecureshare.com/Developers) that exposes workfile read/write, webhook subscriptions, and image attachment endpoints. All fleet agents touching a CCC workfile use Secure Share — no scraping, no screen-reading.
Anthropic's platform for scheduled, API-triggered, and webhook-triggered agent workflows. Runs on Anthropic's cloud infrastructure with built-in audit logging and credential handling. The fleet runs on Routines — nothing is installed on the shop's network.
The industry-recognized reference defining teardown as a billable diagnostic operation separate from repair labor. Used by Agent 04 to attach procedural justification to every teardown line item.
Carrier-issued performance score tied to DRP standing. Delayed or unclear customer communication is one of the top CSI drags — which is why Agent 03's customer SMS output matters even though the direct labor savings are modest.
The industry's clearinghouse for inquiries and rulings on what the estimating systems do and don't include by default. When a shop needs to justify a not-included operation, DEG is the authoritative citation. Agent 01 pulls current DEG rulings on every fire.
Contractual arrangement between a carrier and a shop in which volume is exchanged for agreed-upon labor rates, cycle-time discipline, and operational standards. Every DRP has its own markup rules, delay-reason code schema, and notification obligations — which is why Agent 02's markup calculation and Agent 03's portal notes are per-DRP configured.
The dollar threshold at which a vehicle is declared a total loss. Varies by state (typically 70–80% of ACV). Crossing E01 ends the repair and begins the settlement. Agent 04's entire purpose is predicting this crossing before teardown.
The live delivery promise inside the shop management system. When ECD changes, downstream obligations cascade — rental extension, carrier portal note, customer SMS. Agent 03 listens to ECD changes as its primary trigger.
The initial damage assessment written in CCC ONE, Mitchell, or Audatex. The estimate is the contract for the repair — everything downstream references it. A supplement amends it when additional damage is discovered.
The carrier's initial record of a claim, often with photos uploaded by the customer or field adjuster. Agent 04 reads FNOL photos when available as a secondary signal for severity prediction before the car even arrives at the shop.
A refinish preparation operation on adjacent panels to achieve paint match without visible transition. A frequently missed P-Page line — one of the highest-frequency targets for Agent 01's scrub.
Vehicles combining an internal combustion engine with an electric drivetrain. Both require high-voltage safety procedures when damage affects the HV system. Agent 05 covers HEV, PHEV, and BEV equally.
Hertz's shop/carrier rental portal, equivalent in function to Enterprise ARMS. Most DRP shops juggle both portals depending on which carrier sent the customer. Agent 03 authenticates into whichever applies.
On EVs and hybrids, any system operating above 60V DC. HV procedures include lock-out/tag-out, isolation verification, 50-foot thermal quarantine, and OEM-specific battery attestations. Agent 05 dispatches the correct checklist per make and model.
The industry's technical-reference database for OEM repair procedures and position statements. Agent 05 cross-references I-CAR RTS on every HV-vehicle intake to build the make/model-specific compliance checklist.
The endpoint family (present in Mitchell and Audatex) for programmatically attaching PDFs and images to a workfile. Agent 02 uses ImageAttachment to post sublet invoices into the repair record on arrival.
Recycled or aftermarket parts specified in lieu of OEM. LKQ designation on an estimate line has downstream labor-time implications that the P-Page scrubber accounts for.
A safety procedure where a high-voltage system is deliberately isolated and the isolation is documented with a physical tag and a digital record. A required step on every HV repair — and a document trail Agent 05 produces automatically.
Anthropic's open standard for connecting language models to real tools and data sources. MCP is the plumbing that lets an agent talk to CCC Secure Share, Mitchell, ARMS, email, and anything else — safely and with audit logging.
Mitchell International's developer integration surface (developer.mitchell.com/apis) for workfile read/write, estimate modification, and status subscriptions. The fleet's primary Mitchell integration path.
A collision group running multiple shops under unified ownership — from two-location operators up to regional and national platforms. The Enterprise tier of the fleet is built for MSOs.
An operation that is not built into a given labor time and must be written as a separate line item to be billable. Masking, weld-through primer, feather-prime-block, ceramic coating removal, OEM rivets, and final test drive are all common not-included operations. Agent 01's core job.
OpenAPI Specification v3 — the machine-readable API description format used by Audatex/Solera for their developer integration surface. Governs how the fleet reads and writes Audatex workfiles.
The vehicle manufacturer — Ford, GM, Toyota, Mercedes-Benz, Tesla, and so on. OEM position statements (published repair procedures, mandated parts, mandated calibrations) are the authoritative reference for how a specific vehicle must be repaired. Agent 01 and Agent 05 both consult OEM data on every fire.
The rule document embedded in every estimating system that dictates which operations are included in a labor time and which must be written as a separate line item. Ignoring the P-Page is how shops leak money on every single estimate. Agent 01 lives here.
A specialized repair process for small dents without refinishing. Commonly sublet to PDR specialists — meaning PDR invoices flow through Agent 02's reconciliation workflow.
Our shorthand for the cumulative time a shop burns keeping external systems in sync — rental portals, carrier portals, customer notifications — every time the internal schedule moves. Agent 03 targets this specifically.
The shop-side record of a repair job, often but not always 1:1 with the carrier's workfile. The RO tracks internal scheduling, tech assignment, parts ordering, and cycle time.
In this hub, a routine is a Claude Code Routine — a trigger + prompt + toolset that executes when its event fires. Every agent in the fleet is a routine with a specific configuration.
In this hub, SMS means Shop Management System — the software of record for the repair order, scheduling, and cycle time (CCC ONE, Mitchell, and so on). SMS also means Short Message Service (text messages to customers); context disambiguates.
Work performed off-site by a specialist vendor — scans, calibrations, ADAS work, paintless dent repair, glass, mechanical — that the shop invoices through the primary repair order. Sublet reconciliation is Agent 02's domain.
An amendment to an estimate after additional damage or required operations are discovered, often post-teardown. Agent 01 scrubs supplements on save just like the original estimate.
The disassembly of damaged components to expose all underlying damage for a complete estimate. Teardown is a billable diagnostic operation (CRIB 188) — but carriers often dispute teardown hours on vehicles that subsequently total. Agent 04 prevents the waste and defends the hours.
A claim outcome where the repair cost crosses the E01 threshold and the carrier settles with the owner instead of authorizing repair. The industry total-loss rate ran 22.8% in Q4 2025.
A specific, observable change in a system of record — an assignment landing, an email arriving, a VIN decoding — that a routine can listen for and act on. In our thesis, the trigger is the product: the model is a commodity, the triggers are the differentiator.
The unique 17-character code on every vehicle. Decoding a VIN reveals make, model, year, drivetrain, and build options — and is the trigger signal for Agent 05's HV detection.
The carrier/platform-side record of a claim and its associated estimate(s), supplements, photos, notes, and status. The workfile is the shared source of truth between shop, carrier, and platform — which is why so many agents read from and write into it.