You filed your claim. You did everything right. You went to the doctor, got treatment, sent in your records. And then the insurance company came back with a number so low it felt like an insult.
That wasn't a mistake. That was an algorithm.
Most people don't know this, but the person "reviewing" your claim isn't really reviewing it at all. They're feeding your information into software that was designed, from the ground up, to pay you as little as possible. And if you don't understand how that software works, you're negotiating blind.
I'm going to show you exactly how it works.
Major insurance carriers use claims evaluation software to process injury claims. The most well-known is called Colossus, built by a company called Computer Sciences Corporation. Allstate was one of the first to adopt it in the mid-1990s. Today, virtually every major carrier uses some version of algorithmic claims processing.
Here's what that means for you: when an adjuster enters your medical records, diagnosis codes, and treatment history into the system, the software spits out a valuation range. That range becomes the starting point for your settlement offer.
The adjuster doesn't get to exercise independent judgment. They work within the range the algorithm gives them. If the software says your claim is worth $8,000.00 to $12,000.00, that's the sandbox they play in.
The problem? These systems are built to minimize payouts, not to accurately compensate injured people.
Here's what the software actually does, and why it almost always works against you:
The algorithm assigns different values to different medical treatments. Chiropractic care? Low value. Physical therapy at certain facilities? Lower. Treatment at an orthopedic surgeon's office? Higher.
This means two people with identical injuries can get wildly different offers based solely on where they went for treatment, not whether the treatment helped.
If you miss two weeks of physical therapy because you couldn't get off work, or because you were in too much pain to drive, the software reads that as a "gap in treatment." And gaps in treatment reduce your claim value. Significantly.
The algorithm doesn't know that you're a single parent working two jobs. It doesn't know that your car was totaled and you had no way to get to appointments. It just sees the gap and downgrades your claim.
Your doctor may diagnose you with a cervical sprain, a lumbar disc bulge, and chronic headaches. But the software assigns values based on ICD-10 diagnosis codes, and not all codes are created equal. Some diagnoses trigger higher value ranges. Others get minimized.
If your doctor uses a less specific code, even if the injury is serious, the algorithm gives you less.
Here's the part that makes my blood boil. These systems are terrible at capturing subjective suffering. They can process medical bills and count treatment visits. But they can't measure the fact that you haven't slept through the night in three months. They can't measure that you can't pick up your kid anymore. They can't measure the anxiety you feel every time you get in a car.
Those things are real. They matter. But the algorithm doesn't weigh them properly because they don't fit neatly into a data field.
It gets worse. Insurance companies aren't just using rule-based software anymore. They're moving to machine learning models that predict claim outcomes.
What does that look like in practice?
These systems analyze thousands of past claims to predict what a new claim will settle for. If the model determines that claims like yours typically settle for a certain range, that becomes your offer, regardless of the specific facts of your case.
Some carriers are also using AI to:
That last one is critical. If the algorithm predicts you'll accept a lowball offer without hiring an attorney, you'll get a lowball offer. Every time.
California has some of the strongest consumer protection laws in the country. Under California Insurance Code Section 790.03, insurers are prohibited from unfair claims practices, including making lowball settlement offers that don't reflect the actual value of a claim.
But here's the catch: proving that an algorithm violated fair claims practices is harder than proving a human adjuster did. The software creates a layer of plausible deniability. The carrier can always say, "We used an objective valuation tool."
"Objective" doesn't mean fair.
I'm not telling you this to scare you. I'm telling you this so you can fight back. Here's how:
The first number is almost always the algorithm's floor. It's designed to anchor the negotiation low. If you accept it, the system worked exactly as intended.
The algorithm rewards consistency. Don't skip appointments. If you need to reschedule, do it promptly. Make sure your providers are documenting your symptoms, limitations, and pain levels at every visit.
This matters more than most people realize. A "cervical strain" and a "cervical disc herniation" trigger different valuation ranges. Ask your doctor to be as specific and accurate as possible in their coding.
That friendly call from the adjuster asking "just a few questions"? Your answers are being entered into the system. If you say the wrong thing, the algorithm uses it against you.
I know how that sounds coming from a lawyer. But here's the reality: the algorithm factors in whether you have legal representation. Claims with attorneys consistently settle higher than claims without them. The system knows that lawyers understand how to challenge algorithmic valuations. Unrepresented claimants don't.
Insurance companies spent billions of dollars building these systems. They didn't do that to pay you fairly. They did it to pay you less, faster, and with less pushback.
Understanding how the system works is the first step to not getting crushed by it.
If you've been in an accident anywhere in the San Fernando Valley, Woodland Hills, or greater Los Angeles area, and you got a settlement offer that felt wrong, trust that instinct. It probably was wrong.
Most major carriers use algorithmic claims valuation tools. The most widely known is Colossus, but many companies have developed proprietary AI systems that analyze medical records, diagnosis codes, and treatment patterns to generate settlement ranges.
In most cases, carriers won't voluntarily disclose algorithmic valuations. However, during litigation in California, this information may be discoverable through formal discovery requests.
Yes. Most claims evaluation systems factor in attorney representation as a variable. Claims with legal representation are typically valued higher because the system accounts for the increased likelihood of litigation and formal demands.
Yes, it's currently legal. However, California regulators are increasingly scrutinizing algorithmic decision-making in insurance. The California Department of Insurance has begun examining whether AI tools comply with existing fair claims practices statutes.
Don't accept the offer. Consult with a personal injury attorney who understands how these systems work. A demand that addresses the specific ways algorithmic valuation undercuts your claim is far more effective than a general negotiation.
Artin Nazaryan is a personal injury attorney at Nazaryan Law, APC in Woodland Hills, California, serving clients throughout the San Fernando Valley and greater Los Angeles. If you have questions about your injury claim, call (818) 900-1888 for a free consultation.
Nazaryan Law Car Accident & Injury Lawyers
601 S Brand Blvd, Suite 301, San Fernando, CA 91340
Phone: (818) 900-1888
Artin has a strong track record of securing substantial compensation for clients in motor vehicle accidents, catastrophic injuries, and complex homeowner insurance claims.
why choose Nazaryan Law
Car Accident & Injury Lawyers
"*" indicates required fields
