What Is an AI Mock Jury? A Trial Lawyer's Guide

The basic concept: seeding a case file, seating a jury

An AI mock jury starts where a traditional one does: with your case file. You feed in the facts, the competing narratives, key exhibits, damages theories, and the arguments each side is likely to make. From there, the simulator seats a jury sized to match your actual venue — six, eight, or twelve jurors — rather than an arbitrary panel size chosen for convenience.

The distinguishing idea is that each simulated juror is not a single averaged response but an independent agent. Each one forms an initial reaction based on the record, then reads what the other jurors argue and decides, on its own, whether to hold its position or move. That structure — independent reasoning followed by exposure to competing views — mirrors the basic architecture of real deliberation more closely than a single aggregated survey score would.

How the simulation actually runs

Once the panel is seated, the simulation proceeds across multiple rounds of deliberation. In the first round, jurors react to the raw case file: which facts they found persuasive, which they found weak or confusing, and where they land on liability and damages before hearing from anyone else. In subsequent rounds, each juror is exposed to the reasoning other panelists offered and updates — or holds firm — accordingly, the same way a real jury foreperson might watch consensus build or fracture over the course of an afternoon.

This iterative structure is what allows the tool to surface things a single-pass survey cannot: which argument actually moved a skeptical juror, which theme fell flat even after multiple rounds of discussion, and where the panel split in ways that suggest a hung jury risk. Because the whole process runs as software rather than a scheduled room booking, you can rerun it with a different opening theme, a reordered set of facts, or a stripped-down damages ask and see how the panel's reasoning shifts — in minutes, not weeks.

What an AI mock jury can tell you

The output is a report built for practical case strategy: a verdict prediction framed as directional rather than definitive, a damages range reflecting where the simulated panel landed, a list of case vulnerabilities the panel flagged on its own, and voir dire targeting that highlights the juror profiles most and least receptive to your themes. For a plaintiff's lawyer deciding whether to file a lawsuit or settle, or a defense lawyer sizing exposure before mediation, that combination of signals is often exactly what's needed to make a resourcing decision.

Because you can run the simulation repeatedly at low incremental cost, it is particularly good at answering comparative questions: does leading with the safety violation or the medical timeline land better, does a specific number in closing help or hurt, which juror characteristics correlate with skepticism toward your damages theory. Those are the questions litigators have always wanted a focus group to answer quickly and cheaply, and rarely could.

What it cannot tell you

An AI mock jury does not predict, guarantee, or forecast the outcome of any real jury or real trial, and it should never be sold or relied upon as if it did. It has no access to your actual jury pool's composition, your judge's rulings on motions in limine, the demeanor of your witnesses on the stand, or the thousand small variables that shape a real deliberation room. It is, at bottom, a simulation of how a plausible panel might reason through the themes and facts you feed it — not a window into a specific future verdict.

JuryFish is an argument-testing tool that simulates how a plausible panel might react to your themes. It is a directional aid — not legal advice and not a prediction of any real jury or verdict. Treat its output the way you would treat a smart, well-prepared associate's read on your case: valuable for stress-testing your theory, not a substitute for your own judgment or for the eventual voir dire and trial record.

AI mock jury vs. traditional mock jury and focus groups

A traditional mock jury or focus group remains valuable, particularly for high-value cases where you want to observe real human demeanor, watch live deliberation dynamics, or test witness credibility in person. But that process typically costs $10,000 to $25,000 per session and takes weeks to schedule, recruit for, and report on — a real constraint when you're deciding whether a case theme is even worth developing, or when a mediation date is three days away.

An AI mock jury is not trying to replace that process for every case; it's solving a different problem: the early and iterative testing that happens before you'd ever justify a five-figure focus group spend. Usage runs on credits called fish food, with refill packs starting at $49 and no per-case or per-seat fee, so a solo practitioner testing a single theme and a firm running dozens of matters through the tool each month both pay only for what they use. Many firms use both tools in sequence — AI simulation to sharpen the theory early, a traditional focus group later to validate it on cases where the stakes justify the spend.

When to run one in the case lifecycle

The most common entry point is early case assessment: before you've invested heavily in experts or discovery, running your working theory through a simulated panel can surface whether your core narrative has a fundamental credibility problem worth knowing about now rather than during mediation prep. It's also useful right before a settlement conference or mediation, where a damages range and a list of panel-identified vulnerabilities give you a sharper basis for your number than gut instinct alone.

Later in the case, as trial approaches, the same tool becomes useful for voir dire preparation and closing argument testing — running variations of your themes to see which juror profiles respond well and which arguments hold up after multiple rounds of simulated deliberation rather than just a first impression.

Frequently asked