claimOS vs Independent Contractor Workflows: What Carrier-Side Tools Score, and What a PA File Needs

Same hail line, same carrier, same street in Round Rock. Two adjusters walk the same roof eleven days apart. The first is an independent adjuster deployed through an IA firm: his file opened when a carrier pressed a button, and it will close when someone at a carrier desk marks it QA Approved. The second is a public adjuster whose file opened when a homeowner signed a fee agreement, and it will not close until the money clears the client's account. Both licenses have the word "adjuster" on them. The software underneath them is built for opposite jobs, and the difference is not cosmetic — it is in the data model, the scoreboard, and the billing engine.

The assignment is the unit of work, and someone else owns it

Call it the independent contractor workflow — the ICW stack — and it becomes easier to see what it actually is. Every system in that stack is organized around one object: the assignment. A carrier creates it. A routing rule sends it to the file handler closest to the loss. It moves through a status ladder that the carrier defines — Customer Contacted, Site Inspected, Estimate Returned, Job Started, QA Approved — and when the last status flips, the work is over.

Verisk's XactAnalysis SP is where most of that ladder lives on the service-provider side. Around it, IA-firm platforms bolt on the business layer: auto-assignment by proximity, automatic acknowledgement letters back to the insurer rep, time-and-expense logging, fee schedules, and payout modules that calculate what each 1099 file handler earned per assignment. It is coherent software. It is also, from the first table in the schema to the last report, built around a claim that belongs to somebody else.

The clearest tell is the client record. In an IA-firm system, "client" means the insurer. Rates, protocols, and standards hang off the carrier profile and flow down into every claim that carrier sends. The policyholder — the person whose ceiling is on the floor — is a contact on the file, reachable through a status portal, not the party the system is built to serve.

A public adjuster's file inverts every one of those relationships. Nobody assigns it. It arrives because a homeowner called after a denial, or a contractor made a referral, or a neighbor two doors down recovered $31,000 more than the first offer and told everyone about it. There is no QA Approved status to reach, because there is no desk on the other end of the workflow waiting to approve anything. There is a carrier to persuade.

The scorecard grades you on how far you strayed from the carrier's number

This is the part worth reading closely, because it is published and checkable, and almost nobody in the software conversation quotes it.

XactAnalysis SP has a Performance Scorecard. It rates a service provider in four categories — Cycle Time, Price List Changes, Estimate Quality, and Reinspection — and the overall score is the average of the four. Two details in the vendor's own documentation matter more than the categories themselves. The weights on each metric, per Verisk's help docs, "are designated by the carrier." And the scorecard is only visible to the service provider at all "if a carrier company grants access to it."

Now read the metrics as a public adjuster.

Under Price List Changes: Estimate Deviation Percent, defined as the percentage difference between the price list amount and the estimate amount. Line Items with Price Changes, an average count per estimate. Percent Estimates with Price Deviation.

Under Estimate Quality: Estimates with Overhead and Profit — the percent of estimates that had overhead or profit applied to them. Estimates with Depreciation. Claims with Estimate Corrections.

Under Reinspection: Deviation Percent, the gap between the original estimate and the reinspection. Reinspection Overwrites and Reinspection Underwrites, averaged per file.

Every one of those is a measurement of distance from the carrier's number. On a carrier-side scorecard, distance is variance to be managed. On a public adjuster's file, distance is the entire job. When the price list is stale for your market, deviating from it is not a defect — it is the work product. When a trade count justifies overhead and profit, applying O&P is not a data point to be tracked as a percentage of your estimates — it is money the policy owes. A public adjuster who somehow scored green across that card would be a public adjuster doing the job badly.

None of this makes the scorecard sinister. It is a quality-assurance tool doing exactly what it was built to do, for the party that built it, and cycle-time discipline is genuinely useful work — the same discipline that shows up in how long carriers really take after a named storm. The problem is only that it is the carrier's ruler, and a ruler is never neutral about what it chooses to measure. Run your firm on a system whose reports were designed around that ruler and you will slowly start optimizing for a score you were never supposed to win.

Two operating systems, two scoreboards

Element	Carrier-side / ICW stack	Public adjuster operations
Unit of work	An assignment the carrier opens and closes	An engagement a client opens by signing a fee agreement
Whose clock the software watches	Days to Customer Contact, Days to QA Approved	Statutory acknowledgment, proof-of-loss, and payment deadlines
Estimate deviation	Scored as variance from the carrier price list	Documented and defended — it is the recovery
Overhead and profit	Tracked as a percent of estimates that applied it	Argued trade by trade when the file supports it
Reinspection	Overwrites and underwrites averaged into a score	An evidence package built to survive the re-inspect
How the work gets paid	Fee schedule, daily rate, or time and expense	Contingency percentage of the settlement, capped by statute
Who controls visibility	The carrier grants access to reporting	The firm owns the record and the archive

Two clocks, running in opposite directions

The ICW stack watches Days to Customer Contact and Days in Queue to QA Approved. Those are carrier service levels. Miss them and a threshold gauge turns red on a dashboard the carrier configured.

Your clocks run the other way. They run against the carrier, and they are set by legislatures, not by service-level agreements: acknowledgment windows, proof-of-loss deadlines, payment periods, the interest and fee exposure a carrier faces when it blows past them. A calendar that fires on the carrier's SLA and stays silent on a statutory payment deadline is not a neutral calendar. It is a calendar built for the other side of the table, doing its job faithfully.

The same asymmetry shows up in evidence. The carrier-side platform wants an estimate returned quickly and cleanly against a price list. Your file wants photographs tagged to a room and an elevation, dated to the weather event, and organized so that when a desk adjuster pushes back you respond with citations instead of a search — the discipline that makes a hail roof survive re-inspection. Same roof. Different artifact. Different system.

The statute already answered the question

There is a reading of all this that says: it's just software, put on a different hat, use the tool. Texas disagrees, in writing.

Texas Insurance Code § 4102.158 says a license holder may not represent an insured on a claim, or charge an insured a fee, while representing the insurance carrier against which that claim is made. The same section bars a public adjuster from participating directly or indirectly in the reconstruction or repair of the property, and from taking remuneration or a financial interest from a repair, salvage, or construction firm that picks up work on a claim the adjuster contracted to adjust. Conflicts of interest are not a matter of etiquette in Chapter 4102. They are a licensing condition.

So when the platform's core object is an assignment with a sender and a receiver, it is modeling a relationship that a public adjuster is not permitted to have on the file in front of them. That is not a UX complaint. It is a data model describing the wrong world.

Billing makes the mismatch concrete. An IA-side billing engine handles daily rates, flat fees, fee schedules keyed to claim size, and 1099 payouts to file handlers. A public adjuster's compensation is a percentage of what the client actually recovers, and the percentage is not yours to choose. Florida Statutes § 626.854 caps a public adjuster's compensation at 20 percent of the claim payment — and at 10 percent for claims arising from an event under a governor's declared state of emergency, for claims made in the year following the declaration. The same statute excludes the policy deductible from the amount a fee can be calculated on. Ask a fee-schedule module where you enter a statutory cap, an emergency-period cap that expires on a date, and a deductible exclusion. There is no field. There was never supposed to be one.

What a system built for your side actually counts

claimOS starts from the claim, not the assignment. First notice of loss comes in from a call, a form, or a referral, and every artifact after it — inspection notes, tagged photographs, scope worksheets, carrier correspondence, supplements, the settlement ledger — writes back to that same record. The file you open in week 14 still shows what was true in hour one.

The deadline engine calendars statutory dates by state, not carrier milestones. Withheld depreciation is tracked against a real list of repaired line items, so recoverable money stops slipping past twelve months while everyone waits for invoices. The AI copilot drafts a demand that quotes the line item it relies on and a status update that references the specific weather event, so you can audit the draft against its sources before you send it. Estimating stays where it is: Xactimate and Symbility keep doing what they do, and their output flows into the file for the downstream work — the argument, which is the part that pays. Carriers can stay on Guidewire, and that mismatch is not your problem to solve; the firm running the recovery needs its own system of record.

If you want the direct feature comparison, that lives on the public adjuster page, and the broader market survey is on the public adjuster software comparison.

The one-line test. Open any tool you are evaluating and ask what it does when a claim is closed. If it wants a QA Approved status, it was built for someone whose job ends when the carrier is satisfied. If it wants a settlement figure, a fee calculation, and an archived evidence package, it was built for yours.

That Round Rock roof settled at $67,400 after a supplement the reinspection missed. On the carrier's scorecard, that supplement is a Reinspection Underwrite — a number that pulls somebody's average down. On the homeowner's file, it is a new roof deck instead of a patch. Nobody on the other side of the table is measuring that. Your system had better be.