The full picture of how courts rule isn’t hidden. It’s scattered across thousands of trial court filings. Most law firm brief banks capture only a fraction of it.
Every litigator knows fraud must be pleaded with specificity.
But what does this rule look like in practice?
In this article, we’ll compare two fraud complaints involving similar allegations. Both were challenged by demurrers before the same judge under the same pleading standard. One survived. The other didn’t.
Neither result is surprising once you see the complaints side by side. But seeing them side by side is the whole point. It’s the comparison that reveals how even minute drafting choices can determine whether a claim survives.
The right comparison
Finding one successful complaint isn’t difficult. Finding another complaint involving the same claim, before the same judge, applying the same legal standard—but reaching a different result—is much harder.
Trellis has indexed more than 38,000 rulings on fraud demurrers across the Los Angeles County Superior Court. These rulings span nearly every area of civil litigation, from commercial, insurance, and labor and employment disputes to consumer, property, securities, and tort matters. Each ruling points to the complaint the court evaluated. Taken together, those pleadings form a litigation brief bank.
A litigation brief bank is only as useful as the comparisons it makes possible. If it’s too small, you can’t make the right comparison. If it’s too large, you can’t find it. The challenge is finding a brief bank that’s large enough to contain the comparison, yet searchable enough to retrieve it.
For this comparison, we used the Trellis Connector for Claude to identify fraud complaints involving the same claim, the same judge, and the same pleading standard. Holding those variables constant lets us focus on what actually mattered: the pleading itself.
Here’s what we found.
Same judge. Same law. Different results.
We begin with two fraud complaints. Both involve an individual alleged to have made false representations. Both were challenged by demurrer before the Hon. Cindy Pánuco in Department 224 of the Los Angeles County Superior Court. Both were evaluated under the pleading standard from Lazar v. Superior Court (1996), which requires plaintiffs to allege the who, what, where, when, and how of each alleged misrepresentation.
Carter v. Zuma Partners F I, LLC (20STCV02913)
David Carter alleges that his partner, Henry Shi, secretly made himself the sole owner of their joint management entity. Carter claims that Shi nevertheless continued to represent them both as equal partners, all the while diverting more than $9 million in partnership profits to entities he controlled.
Plus Management, Inc. v. Justin Cho (23STCV15893)
Plus Management alleges that Justin Cho orchestrated a scheme to take over its multimillion-dollar Nordstrom contract. According to the complaint, Cho secretly recruited its employees, misappropriated its confidential files and vendor information, and then emailed Nordstrom to falsely represent that his company had replaced Plus Management as the vendor.
Why the Carter complaint failed
The Lazar standard requires a plaintiff to plead specific, non-conclusory facts for every element of fraud. But what does that look like on the page? Let’s focus on the affirmative misrepresentation element. Here’s what Carter’s fraud claim alleged:
Defendant Shi represented to plaintiff, both orally and in writing, throughout their dealings that they were 50-50 partners in “Zuma Partners,” which the parties both understood was the management company for the investments in the Portfolio Companies, which meant that they would share equally in the manage [sic] company’s profits from such investments. [¶ 105]
The allegation identifies who made the representation (Shi) and what was allegedly said (the parties were equal partners). But it never anchors that representation to a specific communication. The complaint does not identify when the representation was made, where it occurred, or how it was communicated to Carter.
These are the deficiencies Pánuco, the judge, identified while sustaining Zuma Partners’ demurrer. According to the tentative ruling, the allegation that Shi made representations “orally and in writing throughout their dealings” does not satisfy the particularity requirement. It “does not specify how, when, where, or by what means the representations were made to Plaintiff.” Throughout the parties’ dealings refers to a span of time—not a particular interaction. Orally and in writing identifies channels of communication—not the specific statements on which the fraud claim rests.
The complaint that worked
The corresponding allegation in Plus Management read differently. Here’s how the complaint pleaded it:
Justin Cho and Sunwoo Cho established Major and again registered Major as a vendor. Justin Cho sent multiple emails to Nordstrom employees in charge of giving orders to vendors stating Major would work in place of Root and any orders should be given to Major. Sun Woo Cho also acted as the CFO in charge of Major’s operations with vendors and Nordstrom and also human resource management including the Employees. [¶ 52]
Unlike the Carter complaint, the allegation is more particularized. It’s anchored to specific, identifiable communications, pinpointing who made the representations (Justin Cho), to whom they were made (Nordstrom employees), how they were communicated (multiple emails), and what was said (Major had replaced Root).
More importantly, the allegation gives the court something concrete to evaluate. Instead of referring generally to the parties’ dealings, it identifies a specific series of emails, the people who received them, and the false statement those emails allegedly conveyed. The court isn’t asked to infer which communications are at issue. The complaint identifies them directly.
The familiar formulation from Lazar—how, when, where, to whom, and by what means—isn’t a mechanical checklist. It’s a prompt for thinking about what specificity actually looks like. That will vary from judge to judge. Here, Plus Management shows us exactly what a sufficiently particularized allegation looked like to Pánuco.
Concluding thoughts
Most litigators rely on the brief banks their firms have accumulated over time. Those collections are invaluable, but they represent only a small fraction of the pleadings filed every day in state trial courts.
Trellis extends the traditional law firm brief bank by indexing trial court pleadings alongside the rulings that evaluate them. That makes it possible to compare pleadings at a scale no single firm could assemble on its own. With those comparisons, litigators can see how pleading standards are actually applied in practice—not just stated in case law, but tested against real drafting choices.
Get started with the Trellis Connector for Claude
The Trellis Connector brings Trellis’s state trial court data directly into Claude, letting you research judges, attorneys, cases, motions, and rulings through the questions that matter to you.
1. Connect your account
Click the link above, or go to Settings → Connectors in Claude, and select Trellis. You’ll be prompted to log in and authorize access.
2. Start a chat
Once connected, ask Claude what you want to know — no special syntax required. Just let Claude know you want to use Trellis.
3. Let Claude do the digging
Claude will pull from Trellis’s court records to answer directly in the chat, citing the underlying cases, rulings, or filings so you can verify and go deeper whenever you need to.
FAQ
Fraud claims are subject to a heightened pleading standard. A complaint must allege the specific facts supporting each alleged misrepresentation, including who made it, what was said, when and where it occurred, and how it was communicated. General allegations describing a course of conduct are often insufficient to survive a demurrer.
To see how courts have applied this standard in practice, Trellis lets you search trial court complaints alongside the rulings evaluating them, making it possible to compare complaints that survived with those that did not.
Lazar v. Superior Court (1996) 12 Cal.4th 631 is the leading California case on pleading fraud with particularity. It requires plaintiffs to plead the “who, what, when, where, and how” of each alleged misrepresentation. California courts continue to rely on Lazar when evaluating fraud claims at the pleading stage.
With Trellis, you can go beyond the appellate opinion and see how trial courts have applied the Lazar standard by reviewing rulings alongside the pleadings those courts evaluated.
Focus on specific communications rather than general allegations. Identify who made the representation, what was said, who received it, when it occurred, where it happened or how it was communicated, and why it was false. Complaints that survive demurrer typically describe identifiable events rather than broad patterns of conduct.
Trellis makes it easy to compare real complaints applying the same pleading standard, helping you see what sufficient particularity looks like in practice.
Fraud claims are often dismissed because they lack sufficient factual detail. Courts frequently sustain demurrers when a complaint describes an ongoing relationship or general misconduct without identifying the particular statements that allegedly constituted fraud.
With Trellis, you can compare complaints that survived demurrer with those that did not, alongside the rulings explaining the court’s reasoning.
Looking at successful complaints is one of the fastest ways to understand how courts apply pleading standards in practice. The most useful examples involve similar claims, factual allegations, and legal issues.
Trellis indexes trial court complaints alongside the rulings evaluating them, making it possible to find complaints that survived demurrer and compare them with similar complaints that did not.
The most useful comparisons involve similar causes of action, factual allegations, procedural posture, and, when possible, the same judge. Finding those comparisons can be difficult using a traditional law firm brief bank alone.
Trellis makes these comparisons searchable by indexing trial court pleadings alongside the rulings evaluating them, and the Trellis Connector for Claude lets you find similar filings using natural language.
Published appellate decisions establish the legal standard, but trial judges often differ in how they apply that standard to individual pleadings. Reviewing a judge’s prior rulings can provide valuable insight into what level of specificity they expect.
Trellis lets you search rulings by judge and review the underlying complaints that were evaluated, making it easier to understand how a judge has applied a pleading standard in prior cases.
Not by default. General-purpose AI assistants can explain legal principles and summarize published authority, but they do not have access to most state trial court pleadings or rulings.
If you want Claude to search actual trial court filings, connect it to Trellis using the Trellis Connector for Claude. Once connected, you can ask natural-language questions like, “Find fraud complaints that survived demurrer before Judge X,” or “Show me rulings applying Lazar v. Superior Court.” Claude will search Trellis’s trial court database and cite the underlying complaints, motions, and rulings so you can review the original filings.
Yes, but the quality of the results depends on the underlying data. Without access to trial court filings, AI can generate examples or summarize published authority, but it cannot retrieve the actual pleadings filed in similar cases.
To find real examples, connect Claude to Trellis using the Trellis Connector for Claude. Then ask questions like, “Find complaints alleging fraudulent inducement that survived demurrer,” “Show me sample motions to compel before Judge X,” or “Compare complaints involving similar claims.” Claude will search Trellis’s trial court database and return the underlying filings and rulings, giving you examples drawn from actual litigation.
The strongest comparisons hold important variables constant, such as the cause of action, legal standard, procedural posture, and, when possible, the judge. Comparing complaints side by side makes it easier to identify the drafting choices that distinguish a complaint that survived from one that did not.
To make these comparisons, connect Claude to Trellis using the Trellis Connector for Claude, then ask questions like, “Find two fraud complaints decided by the same judge with different demurrer outcomes,” or “Compare complaints applying the Lazar pleading standard.” Claude will search Trellis’s trial court database, retrieve the relevant complaints and rulings, and help you compare them side by side so you can see how courts applied the pleading standard in practice.