The Triage Imperative: How Automated Intake Is Rewiring Insurance Litigation

For most of the modern insurance era, the first contact with a potential lawsuit looked the same: a notice of loss arrived by phone, fax, or mail; a clerk keyed it into a system; and a queue of human adjusters and panel counsel decided, more or less by intuition and seniority, which matters deserved urgency. That model worked when liability claims grew roughly in step with economic inflation. It does not work in an environment where litigation costs have become the single largest driver of liability claims growth, pushing US liability claims up 57% over the past decade according to the Swiss Re Institute. The clipboard has met its match in automated intake and risk scoring, software that ingests, structures, and grades incoming matters for litigation and bad-faith exposure faster than a human can open the envelope.

The Old Way: Intuition, Backlog, and the Cost of a Slow Read

The legacy claims-and-litigation function was built around scarcity of attention. A first notice of loss (FNOL) was a manual artifact, handwritten or dictated, transcribed by staff, and routed through a chain of approvals before anyone assessed whether the matter carried genuine litigation risk. Conflict checks against existing policyholders, claimants, and counsel were run against siloed databases or, in smaller operations, institutional memory. Risk "scoring" was a senior adjuster's gut feel, applied unevenly across thousands of files.

That approach was tolerable when the tail of catastrophic outcomes was thin. It is dangerous now. The era of the "nuclear verdict", a jury award exceeding $10 million, has made the cost of misreading a claim existential. Marathon Strategies documented 135 nuclear verdicts against corporate defendants in 2024, a 52% jump over the prior year, with the total sum surging 116% to $31.3 billion. The median such verdict climbed to $51 million, up from just $21 million in 2020. Forty-nine of these awards crossed the "thermonuclear" $100 million threshold, and five exceeded $1 billion. A single under-triaged file that should have settled early can now blow through tower after tower of coverage.

The macro picture compounds the urgency. The Swiss Re Institute estimates that social inflation contributed roughly 7 percentage points to US liability claims growth in 2023, with US commercial casualty losses growing at an 11% average annual rate to USD 143 billion, a figure that exceeds the USD 108 billion in global natural catastrophe insured losses that same year. The NAIC warns that because social inflation is so hard to measure and reserve for, insurers face heightened insolvency risk, noting that the historic leading cause of liability-insurer insolvency has been under-reserving. In other words, the slow read is no longer just an efficiency problem, it is a solvency problem.

The Shift: From Notice to Score in Seconds

The present-day answer to this pressure is a category of systems that treat intake as a decision point rather than a data-entry chore. Large language models can now parse a claimant's free-text narrative at FNOL, extract structured facts, run automated conflict checks across policy, party, and counsel records, and assign a preliminary litigation-risk score, all before a human adjuster opens the file. Industry analyses of generative AI in claims describe the same pattern repeatedly: the technology transforms FNOL "from a cumbersome, error-prone process into a streamlined, customer-friendly experience," in the words of a Deloitte analysis of generative AI in insurance.

Adoption is no longer theoretical. A survey by the Association of Corporate Counsel and Everlaw found that active generative-AI use among in-house legal professionals more than doubled in a single year, jumping to 52% in 2025 from 23% in 2024. The Thomson Reuters 2025 Generative AI in Professional Services Report found that 26% of legal organizations were actively using the technology, up from 14% a year earlier, and that 95% of professionals expect it to be central to their workflow within five years. Crucially, the same Thomson Reuters data shows the leading legal use cases remain document review, legal research, and summarization, the exact ingestion and triage tasks at the heart of automated intake.

What It Looks Like Now: A Day in the Modern Claims Queue

In a contemporary, AI-assisted claims and litigation operation, the workflow is unrecognizable from the clipboard era. A notice of loss, whether submitted through a portal, a call transcript, or an email, is parsed on arrival. The system extracts the parties, jurisdiction, injury type, policy limits, and narrative facts; cross-references them against open matters for conflicts; and produces a litigation-risk score that flags features correlated with severe outcomes: a plaintiff-friendly venue, a bodily-injury claim, signs of represented counsel, or fact patterns resembling prior nuclear-verdict cases.

Venue is not a trivial signal. Marathon Strategies found that nuclear verdicts in 2024 landed in 34 states and 77 courts, with state courts producing roughly $20 billion of awards against $11 billion in federal courts. Texas led in verdict count, while product-liability and intellectual-property cases drove the largest dollar totals, per the Risk & Insurance summary of the data. An automated scoring layer can weight these jurisdictional and case-type signals consistently, where a human queue would apply them haphazardly.

Data: Marathon Strategies via Insurance Journal.

The downstream economics are striking. Industry analyses of AI-driven claims automation report cycle-time reductions of 40 to 55% and per-claim cost drops in the range of 30 to 40%, alongside meaningful error reduction. A widely cited 2026 industry guide reports insurers using AI-powered claims automation resolving claims up to 75% faster with 30 to 40% lower costs, per CMARIX, while FNOL-automation case studies describe processing-error reductions near 80% and cost reductions around 30%, per Sonant AI. These figures vary by carrier and line of business, but the direction is unambiguous.

Risk scoring also reaches into the bad-faith and regulatory dimensions that make insurance litigation uniquely perilous. A bad-faith claim turns on the insurer's conduct, delays, inadequate investigation, missed deadlines. Automated intake creates a timestamped, auditable record of every action taken on a file, which cuts both ways: it can demonstrate diligence, or it can expose a pattern. With federal marketplace data indicating roughly 19% of in-network claims were denied in 2025 yet fewer than 0.2% appealed, as summarized in a 2025 to 2026 bad-faith analysis, regulators and plaintiffs' counsel are scrutinizing automated decisioning closely.

The Litigation-Funding Multiplier

One reason early triage now carries so much weight is the rise of third-party litigation funding (TPLF), which capitalizes plaintiffs and allows cases to be pressed harder and longer. The NAIC notes TPLF was a roughly $17 billion global industry as of 2021, with just over half deployed in the United States, and that it contributes to growing loss ratios across excess liability, commercial auto, medical malpractice, and general liability, per the NAIC. Reinsurance analysts project the segment will keep expanding; a TransRe overview cites projections that TPLF could reach $31 billion by 2028 and add as much as $50 billion in costs to the US insurance industry over five years, lifting loss ratios an estimated 4 to 5%.

Sources: Swiss Re Institute; Marathon Strategies; NAIC; Duane Morris via Insurance Journal.

Class actions sharpen the picture further. The law firm Duane Morris reported that aggregate settlements across all areas of litigation and enforcement topped $42 billion in 2024, the third consecutive year above $40 billion, with product-liability and mass-tort matters leading, per Insurance Journal's coverage of the firm's review. When exposure aggregates this way, scoring individual matters is not enough; carriers increasingly need to see correlated risk across a book of business, the kind of pattern detection that manual review simply cannot deliver at scale.

The Next Few Years: From Scoring to Strategy

Over the next three to seven years, automated intake and risk scoring will move from describing risk to recommending action. The trajectory is visible in adoption data: Thomson Reuters found that only 15% of law firm respondents consider generative AI central to their workflow today, but 78% expect it to be within five years, per LawSites' coverage of the report. As models mature, the score on a claim will increasingly come bundled with a recommended litigation strategy, settle early, assign to specialist counsel, or reserve aggressively.

Three forces will shape that future. First, measurement will catch up: the ACC/Everlaw survey found only 16% of in-house teams currently track litigation outcomes relative to cost and just 12% track technology ROI, per Everlaw, a gap that, once closed, will let carriers validate which risk scores actually predicted outcomes. Second, regulation will tighten. The NAIC's Property and Casualty (C) Committee is already monitoring these dynamics, and many federal district courts now require disclosure of litigation-funding agreements, signaling a broader push toward transparency in automated decisioning. Third, the bad-faith frontier will harden: as carriers automate denials and reserve decisions, the same auditable trails that defend diligence will be discoverable evidence of pattern conduct.

The risks are real. An over-tuned scoring model that systematically under-reserves dangerous claims recreates the very under-reserving the NAIC identifies as the leading cause of liability-insurer insolvency. A model that flags claims for fast denial without human judgment invites the bad-faith exposure it was meant to reduce. The carriers that win will treat risk scoring as a decision-support layer, fast, consistent, auditable, rather than an autopilot.

Conclusion

The insurance industry spent a century treating intake as paperwork. The data now make clear it is the most consequential moment in the litigation lifecycle. With liability claims growing far faster than the economy, nuclear verdicts setting records, and litigation funding multiplying the number of cases pressed to trial, carriers can no longer afford a slow, intuitive, inconsistent first read. Automated intake and risk scoring close that gap, turning the notice of loss from a clerical bottleneck into a strategic checkpoint. The technology is generic and the category is young, but the direction is set: the future of insurance litigation belongs to those who can score a claim's risk before the clipboard would have even been filed.

Sources

1. Swiss Re Institute, "Litigation costs drive US liability claims by 57% over past decade." swissre.com
2. Swiss Re Institute, sigma 4/2024: Social Inflation: Litigation Costs Drive Claims Inflation. swissre.com
3. NAIC, Insurance Topics: Social Inflation. content.naic.org
4. Marathon Strategies via Insurance Journal, "Corporate Nuclear Verdicts Surged to New Record High in 2024." insurancejournal.com
5. Risk & Insurance, "Nuclear Verdicts Skyrocket: Corporate Lawsuit Awards Surge 116% to $31.3 Billion in 2024." riskandinsurance.com
6. MoneyGeek, "Nuclear Verdicts and Small Business Insurance." moneygeek.com
7. Association of Corporate Counsel / Everlaw via Business Wire, 2025 GenAI Survey. businesswire.com
8. Everlaw, "81% of CLOs Say GenAI Accelerates Legal Work." everlaw.com
9. Thomson Reuters, 2025 Generative AI in Professional Services Report (executive summary). legal.thomsonreuters.com
10. Thomson Reuters, Press release: GenAI adoption nearly doubles. thomsonreuters.com
11. LawSites, "Over 95% of Legal Professionals Expect Gen AI to Become Central to Workflow Within Five Years." lawnext.com
12. Deloitte, "Transforming customer experience in insurance: Harnessing the power of Generative AI." deloitte.com
13. Insurance Journal, "Settlements Eclipse $40 Billion in 2024." insurancejournal.com
14. Duane Morris Class Action Defense, "Settlement Numbers Break $40 Billion For The Third Year In A Row." blogs.duanemorris.com
15. Legal Dive, "Plaintiffs reap $160B in 3-year class action haul." legaldive.com
16. TransRe, "Social Inflation Overview 2025." transre.com
17. CMARIX, "AI in Insurance Claims Processing: 2026 Automation Guide." cmarix.com
18. Sonant AI, "FNOL Automation: Transform Claims Processing in 2026." sonant.ai
19. Impactia, "AI Claims Processing Automation for Insurance." impactia.ai
20. Best Attorney USA, "Comprehensive Analysis of Insurance Bad Faith, 2025 to 2026." bestattorneyus.com