Ours is our own internal test. The Lexis and Westlaw figures are from Stanford’s own separate published research, not a controlled side by side test against us.
Two rules do the work.
Show the source, or stay silent.
Every answer comes with the exact passage from the real document it relied on. If the system cannot find a passage that supports a point, it says so instead of filling the gap. A fabricated case has no passage to show, so it has nowhere to hide.
Check whether the case is still good law.
Before any authority reaches you, the system checks whether a later court overruled, criticized, distinguished, or set it aside, and it flags the case with the reason. This is a signal to guide your review, not a substitute for a Shepard’s or KeyCite check.
It reads their citations too.
Below is a passage from an answering brief in a sample matter. Click either citation and see the page it actually points to.
One of these two holds up. On a recent Nevada appeal this found several that did not.
A true quote can still lose you the argument.
The passage below is real, quoted correctly, and the citation is right.
Counsel emphasises the child’s school stability, and the record supports the fact. It does not support the inference drawn from it.
The child has been enrolled in the Clark County district continuously since August 2023.
That continuity has been maintained principally through Father’s residence three blocks from the campus and his assumption of morning transportation since the separation.
Same sentence. Read alone it supports your client. Read in place it supports theirs. No citation checker catches this, because nothing about the citation is wrong.
A hallucination is when an AI states a case, quote, or rule that does not exist or does not say what the AI claims. A fabricated citation in a brief is a problem of candor with the court, not a small typo.
Most AI answers from memory, and that is where invented cases come from. DilloLex first pulls up the actual documents that relate to your question, then answers only from those documents. Reading real text instead of guessing from memory is the single biggest reason fabricated answers drop.
Finding the right document takes two steps. The first casts a wide net, fast. The second reads that set with more care and keeps only the documents that truly answer the question. A tool that skips the second pass serves documents that sound related but are not.
No. Our testing was done by our own team on a Nevada question set, so the numbers are our own honest measurements, not certified by an outside party. Ask us about our method at a demo, and we will walk you and your malpractice carrier through it.
We have not run a controlled side by side test against the big research services on the same questions, and we will not pretend we have. What we claim is a process you can defend: answers built from real documents, a source shown for every point, a good law check on every case, and your review before it leaves the firm.
See it for yourself. Bring your questions.
If you need to take it to your malpractice carrier, we can do that too.
Every day, 8am to 8pm. A founder answers — never a salesperson.