Know the other side’s expert first, and recover about 10% more
The side that anticipates the opposing expert wins the expert fight more often, and the expert fight decides a large share of high-stakes cases. Across the seven practice areas on our 2027 roadmap, preparing with Supreme Mind is worth about +9.8% in expected recovery against a firm without AI, and +16.5% as the agents improve.
What it is worth on one case
Expected recovery on one close matter in each practice area, when the plaintiff side prepares with each option. Securities and antitrust are class-certification fights from our research model; the others are illustrative matters.
Two experts, one case, two answersSee where your expert and the defense expert split, in a typical matter in each practice area
Both sides hire the same kind of expert and usually the same method. The reports still land far apart, because each expert makes five or six judgment calls inside that method, and each call has a defensible range. Typical matters, with illustrative figures.
Indirect-purchaser cases add one more split, pass-through: the plaintiff’s expert puts it at 80 to 100%; the defense expert says it is too uneven to measure across the class.
Illustrative matters built from typical filings. Not a prediction for any specific case.
Across every practice area we focus on
Increase in expected recovery against a firm without AI.
Your expert is the one most often challenged
Published rates, when the expert fight is fought. They are why anticipating the other side’s expert, and stress-testing your own, moves outcomes.
Because it gets the expert’s analysis right
Share of the opposing expert’s key findings anticipated before the report is served. Securities is measured on our blind benchmark; the other areas are estimates.
Where in a matter the value is made
Share of the gain by moment, and how often a team runs it. Case selection matters: about 55% of securities class actions settle before a certification motion, and funders back fewer than 5% of the cases they review.
Intake and case selection
25%1 to 2 runsRepeated runs as facts develop
25%10 to 20 runsClass certification and Daubert
30%3 to 5 runsRebuttal and cross
10%2 to 4 runsMediation and settlement
10%1 to 2 runsSee what it finds on one of your own matters
Modeled estimates, not measured outcomes. Securities and antitrust figures for Supreme Mind, on both sides, come from the outcome model on Research, for one side preparing in a close, contested matter. The other practice areas are scaled by how much the expert decides the outcome there, using published exclusion and outcome data (PwC 2022; Charles River Associates 2025; Peruzzi 2024; a 2015 to 2020 study of IP damages rulings, in which lost-profits opinions were excluded 14% of the time; Cornerstone Research on Delaware appraisal, where premiums fell from +27% before 2017 to −8% since), and general AI and the ceiling by their accuracy gain over a firm without AI. Accuracy for securities is measured on the Benchmark; all other accuracy figures and stage shares are estimates. No public source reports expert spend by practice area, so the weights are our estimate. Settlement timing: Cornerstone Research, 2025 Review (February 2026). Funder selectivity: GAO (December 2022) and Burford Capital. Defense values were checked against the Searle Civil Justice Institute (2015), CRA (2025) and mass-tort outcomes, and are discounted for appeal risk: in July 2026 the Second Circuit reversed three of five general-causation exclusions in the acetaminophen litigation. Better preparation can also lower defense costs, which this page does not count. The case examples are close, contested matters; the typical-case range applies real-world certification and exclusion rates and a preparation effect from half the assumed improvement to its low end, which gives about 40% to 75% of the close-case gain. Measured results from the pilots will replace these figures.
No guarantee of any result. These figures illustrate what better preparation can be worth across many matters; they are not a prediction or promise for any particular case. Supreme Mind provides decision-support materials. It does not give legal advice, act as an expert witness or represent any party, and counsel of record decides whether and how to use any output. The outcome of every matter depends on its facts, the law, the court, the parties and counsel’s own judgment, so Supreme Mind is not responsible for the outcome of any matter in which its Expert Agents are used. See our Terms.