Six minutes of lecture
The plaintiff's expert is a structural engineer with twenty-five years of experience, three textbooks and a calm, patient manner. On direct she explained, clearly and persuasively, why the building's collapse was caused by a design defect.
The associate conducting the cross has read her report twice and consulted the defense expert. His first question is about her methodology.
"Isn't it true that finite element analysis has significant limitations?"
She smiles. "That's a good question, and it's one I'm often asked. Let me explain..."
Six minutes later she has delivered a clear, accurate and persuasive lecture to the jury about finite element analysis, its limitations, why they don't apply here, and why her conclusions are sound. The associate's first question has become her second direct examination.
Expertise is her territory
Cross-examining an expert witness is a distinct discipline from ordinary cross, because the usual dynamic reverses. With a lay witness, the lawyer usually knows more about the case than the witness does. With an expert, the witness usually knows more about the subject than the lawyer ever will.
Three features make this dangerous.
Open questions are gifts. Experts are skilled explainers and many testify frequently. Any question that allows explanation will receive a persuasive one.
Technical challenges are unwinnable. A lawyer who argues science with a scientist will lose — not because the lawyer is wrong, necessarily, but because the jury will believe the person with the credentials.
And the jury likes experts who teach. A calm expert explaining patiently in response to aggressive questioning looks more credible, not less.
The four targets
Effective expert cross rarely attacks the expert's field. It attacks what the expert did with it.
Assumptions. Every expert opinion rests on assumptions — about facts, measurements, sequence. "Your opinion assumes the load on the third floor was as shown in the original plans." "Yes." "You didn't measure the actual load." "No." "If the actual load was higher, your calculation would change." "It could." The expert's conclusion is only as good as the assumption, and the assumption is usually something the expert didn't verify.
Inputs. What data did the expert review — and what didn't she? "You reviewed the design drawings." "Yes." "You didn't review the maintenance records for 2019 to 2022." "They weren't provided to me." The point isn't that she was careless; it's that her opinion was formed on an incomplete record.
Scope. Experts often testify slightly beyond the edge of their expertise. "You're a structural engineer." "Yes." "You're not a materials scientist." "No." "The corrosion analysis in your report came from someone else." "It was provided to me, yes."
Consistency. Prior testimony, publications and earlier reports. If the expert has said something different in another case or in her own writing, that inconsistency is often the strongest material available — and it must be used precisely and accurately.
The technique is the same — only stricter
The single-fact, leading question discipline covered earlier in this series applies with more force to experts, because the cost of an open door is higher.
Never ask a question she can explain. "Isn't it true that..." followed by a contested proposition invites a lecture.
Ask about what she did, not what she knows. Facts about her process — what she reviewed, measured, assumed, relied on — are hard to lecture about.
Get concessions, not victories. The aim of expert cross is usually a series of small, accurate concessions that the other side's expert or the closing argument can then use. Rarely does an expert "break" on cross, and trying to make her do so often backfires.
Don't ask her to agree with your conclusion. She won't, and she'll explain why.
Be accurate about her prior statements. Mischaracterizing an expert's publication or earlier testimony in front of a jury, with the expert present to correct it, is one of the fastest ways to lose credibility.
Compensation and bias — carefully
Expert fees and the frequency of testifying for one side are often legitimate subjects of cross, subject to the rules and practices of the court. They can be effective, and they can backfire when handled with sarcasm or when the lawyer's own expert is paid similarly.
Ask plainly, briefly, and move on. The fact does the work; the tone shouldn't.
Four ways it goes wrong
The science debater, who challenges methodology in open questions and gets a lecture.
The overreacher, who tries to make the expert concede her conclusion is wrong.
The mischaracterizer, who misstates the expert's prior writing or testimony and is corrected in front of the jury.
The unprepared examiner, who hasn't identified the assumptions and inputs the opinion depends on, and has nothing to cross on except the science.
Why this isn't trained
Associates practice cross on lay witnesses. Expert cross is usually handled by senior lawyers, so juniors get little practice before they're asked to do it.
Preparation focuses on content. Associates spend hours with the defense expert learning the science — and then use that knowledge to argue it, which is the wrong use.
Real experts are expensive to rehearse against. Firms rarely have an expert available to play the opposing expert in practice sessions.
And peer role play produces a weak expert. A colleague pretending to be an engineer can't deliver a persuasive, calm, technically accurate lecture in response to an open question — which is the exact thing the associate needs to experience.
What expert cross-examination simulation can rehearse
A ten-minute simulation can put the lawyer opposite a patient, articulate expert who seizes every open question and responds precisely to single-fact questions about assumptions, inputs, scope and consistency. The AI agent in Foretell AI plays the expert using the matter's report and materials; the substantive analysis, the firm's own expert input and trial strategy stay with the firm.
Four versions to build:
- The lecturer, who turns open questions into explanations.
- The assumption witness, whose opinion depends on an unverified factual premise.
- The scope-stretcher, who testifies beyond her specialty.
- The prior-statement expert, whose earlier writing is partially inconsistent — testing whether the examiner uses it accurately.
Design caution. Rules on expert testimony, the admissibility of opinion evidence, use of learned treatises and the permissible scope of cross vary between federal and state courts and across jurisdictions. Scenarios must use the matter's actual reports and materials, and technical content should be reviewed by the firm's own experts; nothing here is legal or technical advice.
Designing the module
Ten minutes, scored against an expert cross rubric focused on target selection and question discipline.
Pass one — target selection. How many questions addressed assumptions, inputs, scope or consistency — versus the underlying science?
Pass two — question design. What proportion were single-fact leading questions? How many invited explanation?
Pass three — accuracy and restraint. Were prior statements characterized accurately? Did the examiner avoid asking for the ultimate concession?
Rubric on observable behavior: Proportion of questions on process rather than science. Proportion of single-fact leading questions. Number of explanations elicited. Were concessions obtained on assumptions or inputs? Were prior statements used accurately? Was the ultimate conclusion challenged directly?
Explanation count is the measure. Every explanation is the expert's opportunity; the fewer the examiner gives her, the better the cross.
For litigation firms and legal departments
Expert testimony often decides complex cases. Product liability, construction, IP, financial disputes — the expert battle is frequently the case.
Juniors are rarely given practice. Simulation lets associates develop expert cross before they're asked to do it for a client.
It improves preparation with your own expert. Associates who know the cross targets ask their own experts better questions: what assumptions does the other side rely on, what data didn't they review?
And for law schools, advanced trial advocacy courses can add expert witnesses without the cost of recruiting real experts.
Frequently asked questions
How do you cross-examine an expert witness? Focus on the expert's assumptions, the data they relied on or didn't review, the limits of their expertise, and any inconsistencies with their prior statements — using short, single-fact leading questions.
Should you challenge an expert's methodology on cross? Rarely directly, and never with open questions. Challenging the science invites a persuasive explanation. Challenging what the expert assumed or didn't review is usually more effective.
What is the goal of expert cross-examination? Usually a series of accurate, limited concessions that undermine the foundation of the opinion — not a dramatic reversal.
Can you cross-examine an expert about their fees? Often, subject to the court's rules and practice. Keep it brief and factual; sarcasm tends to backfire.
The short version
He asked her about her methodology and she gave the jury a six-minute seminar. That was his fault, not hers.
Don't argue the science. Cross the assumptions, the inputs, the scope and the consistency. Ask about what she did, not what she knows. Single-fact, leading, short. Collect small concessions and leave them for closing. Quote her prior statements precisely or not at all.
You won't out-know an expert. You can show the jury what her opinion stands on.
Foretell AI lets litigation firms and law schools build expert witness simulations — including lecturing experts, assumption-based opinions and prior-statement cross like the one above — with configurable AI experts, recordings and rubric-based evaluation. If your associates first cross an expert with a real jury watching, we're happy to walk through how other firms have structured it.