The first claim looked routine. A low-speed rear-end collision at an intersection, four occupants, soft-tissue injuries, treatment at a chiropractic clinic across town, a demand letter from counsel six weeks later. The adjuster paid it.
Eighteen months later, the special investigations unit is looking at forty-one bodily injury and PIP claims that share almost nothing on their face. Different insureds. Different vehicles. Different intersections. Different policies. But the same clinic shows up on thirty of them. The same tow company on twenty-two. The same law firm sent demand letters on twenty-six. And one name, a passenger in a 2025 accident, appears again as a driver in a 2026 accident and as a witness in a third.
Carrier counsel is now weighing a civil RICO complaint against the clinic and its owners, rescission on a handful of policies, and a referral to the state fraud bureau. Every one of those paths needs the same thing first: proof that these are not forty-one accidents but one operation.
What the claim file actually contains
An organized fraud ring case, whether it ends in a denial, a declaratory judgment action, or an affirmative suit against providers, is built from records the carrier already holds or can get:
- First notice of loss reports and claim notes, exported from the claims system as spreadsheets
- Policy applications with addresses, phone numbers, garaging locations, and prior insurance
- Medical bills and treatment records, with provider names, tax IDs, and dates of service
- Tow, body shop, and rental invoices
- Transcripts from examinations under oath (EUOs) and recorded statements
- Police crash reports listing drivers, passengers, and witnesses
- Phone numbers from the claims, and where subpoenaed, call detail records
Reviewed one at a time; each claim can look clean. The ring only exists in the overlap between them.
Where the tabs stop working
The usual approach is a master spreadsheet. One row per claim. Columns for claimant, clinic, attorney, tow company, date of loss, location. Sort by clinic, count the hits, highlight the repeats.
That finds the obvious pattern: one clinic, thirty claims. It doesn’t find the rest. The passenger who became a driver used a different address the second time. The phone number on one claimant’s application belongs to the clinic’s office manager. Two “unrelated” insureds list the same garaging address. The tow company’s owner shares a last name and a mailing address with the clinic’s billing contact. The recruiter who signs up participants never appears on a claim form, but his number shows up in the call records of nine claimants in the days before their accidents.
Those are the connections that turn a suspicious pattern into a provable enterprise. In a spreadsheet, each one is a lookup somebody has to think to run, across tabs, against names spelled three different ways.
In SentVi, every claim, person, vehicle, provider, address, phone number, and business is an entity. Every link between them (treated at, towed by, represented by, called, lived at, rode with) is a relationship with a date attached. Import the claims export, the billing data, and the crash reports, and the ring draws itself. Forty-one separate claims become one network with a handful of people at its center.
The three views that matter in a fraud ring case
The network. Social network analysis ranks entities by how connected they are and by whether they sit between otherwise separate groups. In a staged-accident ring, the claimants are usually the least important people on the chart. The people who score highest on betweenness are the ones connecting claimants to the clinic and the lawyers: the recruiter, the office manager, the owner’s relative who shows up on two tow invoices and never on a claim.
The timeline. Rings run on a rhythm. Policies bound shortly before a loss. Accidents clustered on the same days of the week. Treatment that starts the day after the collision and stops when the PIP limit is reached. SentVi’s temporal view places every policy inception, loss date, date of service, and call on a single axis, and the time-range slider shows how the network grew, claim by claim. That pattern is hard to explain as coincidence in an EUO or a deposition.
The map. Staged accidents repeat locations. When eleven losses occur at four intersections, when claimants living across a metro area all treat at one clinic, when garaging addresses cluster on a single block, the geospatial view shows it at a glance and exports to Google Earth for investigators in the field.
From the chart to the complaint
Fraud ring litigation is heavy on documents and heavy on skeptics. A judge or jury has to see an enterprise, not a list. SentVi exports network diagrams, timelines, and maps as exhibits anyone can read without training. SentVi Reader lets outside counsel, the SIU, and a forensic expert work from the same interactive chart without each needing a license. When the next batch of claims arrives, it goes into the same chart, and new participants either connect to the ring or they don’t.
Because SentVi runs on your own hardware and can operate fully disconnected from the internet, claimant medical records and policyholder data stay under your control.
The question the investigation should ask
Not “which claims look suspicious,” but “who connects them.” A spreadsheet can count how often a clinic appears. Link analysis shows who put the claimants in the car.
See the workflow in the video below, or request a demo: https://sentvi.com/link-analysis-book-a-demo/
