When you submit a personal injury claim, an adjuster doesn’t just “consider your case.” They feed it into proprietary software that outputs a settlement number. Understanding which software each insurer uses, and what inputs drive the output, is how you negotiate against the algorithm instead of getting steamrolled by it.<\/p>\n
Used by:<\/strong> Allstate, Liberty Mutual, AIG, USAA, MetLife, Aviva, and 50+ other carriers. Roughly 60% of U.S. auto BI claims pass through Colossus.<\/p>\n How it works:<\/strong> Adjusters input 720+ fields about your case \u2014 injury codes, treatment dates, age, occupation, jurisdiction, attorney involvement, MMI status, etc. Colossus outputs a “Severity Point Value” which translates to a settlement range.<\/p>\n Known weakness:<\/strong> Colossus heavily discounts subjective complaints (pain levels, emotional distress) without objective imaging or clinical findings. It also undervalues older injuries (historically treats >6-month claims as “stale”).<\/p>\n Used by:<\/strong> GEICO, Progressive, Farmers, Nationwide, MetLife (some lines).<\/p>\n How it works:<\/strong> Similar input model to Colossus. ClaimIQ also pulls comparable settlement data from a proprietary database of claims handled within the Mitchell network \u2014 which is itself biased low because most settled claims were undervalued.<\/p>\n Known weakness:<\/strong> Heavy reliance on ICD-10 injury codes. If your treating doctor used a low-severity code (common error), ClaimIQ tags your claim as low-severity even if the actual injury was significant.<\/p>\n State Farm:<\/strong> uses an internal “ACE” (Advanced Claims Evaluation) system, the details of which State Farm has fought aggressively to keep secret.<\/p>\n Farmers:<\/strong> developed an in-house system after moving away from Colossus.<\/p>\n Smaller carriers:<\/strong> often use spreadsheet-based valuation tools developed by the corporate claims department.<\/p>\n Whatever system the insurer uses, the same variables move the output number. In approximate order of impact:<\/p>\n Adjusters don’t simply pay what the algorithm calculates. They have authority bands:<\/p>\n This is why “final” offers move 30-50% when you push. The line adjuster’s “final” is below the supervisor’s ceiling.<\/p>\n Algorithms can’t see what isn’t in the file. Submit:<\/p>\n Each of these forces the algorithm to recalculate higher.<\/p>\n Request your medical records and review the diagnostic codes. If your “cervical strain” should have been “cervical disc displacement,” ask your doctor to amend. The algorithm reads codes, not narratives.<\/p>\n Insurance algorithms calibrate to settlement averages, not jury verdicts. Citing comparable jury awards in your state ($100K+ over the algorithm output) signals that going to trial threatens a worse outcome than paying you fairly now.<\/p>\n A handful of states (NJ, Mississippi cases) have allowed plaintiffs to depose Colossus details in litigation. The threat of forcing this discovery often unsticks negotiations.<\/p>\n2. Mitchell ClaimIQ (formerly Decision Point)<\/h3>\n
3. Proprietary Insurer Systems<\/h3>\n
The 12 Variables That Drive Insurance Algorithms<\/h2>\n
\n
Why You’re Always Getting a Lowball First Offer<\/h2>\n
\n
How to Argue Against the Algorithm<\/h2>\n
Tactic 1: Boost the Inputs<\/h3>\n
\n
Tactic 2: Challenge the ICD-10 Coding<\/h3>\n
Tactic 3: Reference Jury Verdicts<\/h3>\n
Tactic 4: Demand Disclosure (Some States)<\/h3>\n
Carrier-Specific Patterns<\/h2>\n