The standard answer has become inexpensive
The machine types out the 90% that repeat themselves — reliably and in seconds. What used to take hours now costs almost nothing. Which is exactly why it is no longer what you need a law firm for.
You have an AI result.
We do the last 10%.
Give us your result. We check not only whether it looks good — but whether your case is the typical case it was made for. We do not dismiss it, and we don't start from scratch. We build on it.
The machine is good — until it really matters.
[ When it counts ]It follows the most traveled path
A language model chooses the most probable path — the one most cases take. Like traffic following the widest road: usually right, and wrong precisely when your case is not part of the majority.
Those cases end up with us
Where the typical solution and the right solution diverge, the machine stops being enough — and that is where lawyers' work begins. AI now resolves the simple cases. What remains are the difficult ones.
You can't tell by looking at the result
Whether an AI result holds up or leads you astray is not visible from the wording alone — it sounds confident either way. Whether your case is the typical case can only be answered from the outside: by someone who recognizes the outlier.
The starting point
The last 10% are not the remainder. They are the whole job.
What we do with your result.
[ Three paths, depending on your case ]A quick assessment
We examine your result and clarify two points: whether the output is usable at all — and which counterarguments can be raised against it, including the facts the result tacitly assumes. We do this on the basis that we accept your account of the facts as true, without verifying it.
Typical case or outlier
The real question. We assess whether your case is the typical case the machine was made for — or the outlier where its answer is systematically off. That is the information your result itself does not provide, and it determines everything that follows.
If your case falls outside the pattern
If your case falls outside the pattern, we build on the AI result: the machine has delivered the typical case, you provide the facts and your objective, and we provide the legal judgment — and the responsibility for it.
Which stage fits your case is something we clarify in the initial contact — as is the scope.
What we review — and what we don'tNo review should
No review should
promise more than it can see.
An AI quick check is not a fact-finding exercise. We take your facts as they appear in your AI result and review the result on that basis. If the basis is wrong, the result changes.
That is not fine print — it is the reason the quick check can be lean, and the reason an outlier needs more than a glance at the output. Anyone who discloses this is not selling you false certainty.
Before you submit.
[ Frequently asked questions ]Does submitting already create a lawyer-client relationship?
No. Your submission is an inquiry. A lawyer-client relationship or engagement agreement (Mandatsvertrag) is created only by express agreement. Until then, we treat your information confidentially and first check whether a conflict of interest exists.
How and where is my data processed?
Your information is processed in the EU. We ask that you refrain from including real names of third parties and particularly sensitive details in the initial contact — everything else can be addressed within the protected lawyer-client relationship.
Which AI tools do you accept?
All of them. Whether ChatGPT, Copilot, an industry tool, or a system you built yourself — we care about the result and your case, not the tool.
Do you verify whether the AI result is “correct”?
Not conclusively in the quick check. We tell you whether a usable result exists and which counterarguments there are — based on the facts as they appear in your AI result. A reliable legal assessment requires follow-up legal advice.
What happens if it is an outlier?
Then we tell you and propose how to build on the result. You decide whether and how to proceed.
How does this differ from the initial assessment?
The initial assessment is an open conversation about your matter. The AI second opinion starts from a concrete, already existing AI result. The two can be combined.
Give us your result.
[ Submit ]Briefly describe the matter and paste your AI result. We first check whether a conflict of interest exists, and then get back to you.
Prefer to
talk first?
The AI second opinion does not replace a conversation — it sharpens it. If you would rather talk directly: initial assessment, 15 minutes, free of charge, protected by lawyer-client confidentiality.
- Firm
- JUN Legal GmbH · Rechtsanwaltsgesellschaft
- Address
- Salvatorstraße 21 · 97074 Würzburg
- Phone
- +49 931 6639232
- info@jun.legal