Every case has a hole. Find yours first.

There is a particular silence that falls in a courtroom when the bench asks a question nobody on your team had considered.

Every litigator in India knows that silence. You have read the brief four times. Your junior has tabbed the compilation. You have gone over the opening in the car, in the lift, in the corridor outside the court. And then, somewhere in the second paragraph of your submission, a judge looks up from the file and asks one short question that goes directly to the weakest part of your case.

You recover. You seek a short accommodation, you offer to address it after the break, you find something in the record. But you know what has happened. The case did not fail. You met your own weakness in front of the bench rather than in your chamber.

That moment is what we set out to move. Not to remove it, but to move it earlier.

After years spent around courts, legal data, and the people who practise law, one thing has become clear to me.

Lawyers are rarely blind to the weaknesses in their own cases. Ask a competent advocate in an unguarded moment and they will tell you precisely which paragraph of the petition concerns them, which document in the annexures they wish were not on record, which precedent they hope the other side has not read closely.

Awareness was never the difficulty. The difficulty is that the awareness remains untested until something forces it into the open, and in litigation that something is usually opposing counsel or the bench, at the least convenient moment available.

What a case needs before it reaches court is not agreement. It needs opposition.

The devil's advocate is not a modern invention. It is a formal role of long standing, created for a single purpose: to argue against a proposition so that its weaknesses are exposed before a decision is taken. A role defined not by helpfulness but by resistance.

Every serious argument needs one. Very few lawyers have access to one.

The reasons are practical.

The ablest junior in the chamber is already carrying three other matters and the day's cause list. The senior whose view you would most value bills by the hour and is unlikely to sit through a two hour session on a matter they are not appearing in. Co-counsel would prefer the conference to conclude. And in most chambers there is a hierarchy in the room, which means the person best placed to identify the flaw is often the person least placed to say so.

The result is that a conference intended as a stress test becomes a rehearsal. Positions are affirmed rather than examined. The weak point survives, undisturbed, until the hearing.

Software is indifferent to that hierarchy. It has no relationship to preserve and no incentive to soften an assessment. That indifference is not a small feature. It is the reason the role can finally be filled reliably.

That is the gap Devil's Advocate was built to close.

Devil's Advocate sits within MIRA, our reasoning layer. It works in three stages.

First, you choose your side.

Petitioner or respondent. Plaintiff or defendant. You set the forum, the jurisdiction, and the practice area. A writ before a High Court is not a commercial suit, and neither is a tribunal appeal. An argument that succeeds in one forum may not survive in another, so the frame is established at the outset, by you.

Second, you argue.

Turn by turn. You advance your case and MIRA takes the opposing position, in the manner opposing counsel would.

This is the part that required the most work and mattered most to get right. It rebuts your strongest submissions with real, cited case law. It concedes what is genuinely sound, because a capable opponent does the same, and an opponent who contests every point teaches you nothing about where the real risk lies. It presses hardest where the argument is thinnest.

No strawmen. No contrarianism for its own sake. An adversary that disagrees indiscriminately provides no useful signal, and experienced counsel will recognise that immediately.

Third, when you are ready, you ask the judge.

An impartial AI bench delivers a reasoned verdict. It frames the issues. It sets out the strengths and weaknesses of each side. It states the likely outcome on the arguments as they currently stand, and identifies what would need to be strengthened before the hearing.

That final element is the most useful output of the exercise. It is not a prediction of the result. It is a list of what requires attention.

I want to be precise on this point, because the Indian legal market has sound reasons to be cautious about systems that speak with confidence.

An opponent that invents authority is worse than no opponent at all. If MIRA rebuts you with a judgment that does not exist, or attributes to a judgment a proposition it never laid down, we have not spared you a difficult hearing. We have created one.

Devil's Advocate therefore argues from our verified legal corpus, with citations attached and traceable, and with the same verification and cross-validation layer that sits above every other output in MIRA. Reasoning that can be explained, sources that can be checked, and opposing views set out rather than smoothed over.

This is also why the work sits with us. Our family's engagement with this profession began in 1937, when Advocate Ganesh Lal Mandal established his practice. My father, S. C. Yadav, spent his career structuring legal knowledge long before it was described as data. I went on to lead the Central Data Processing Centre of the National Judicial Reference System, which built one of the largest government repositories of legal appeals in the world.

The conclusion drawn from three generations of that work is straightforward. In law, reasoning is only as reliable as the record on which it stands.

It is not a substitute for professional judgement. The decision on strategy remains yours, and so does the decision that matters in court.

It does not forecast outcomes. It assesses how an argument withstands competent opposition in the form it has presently taken. Revise the argument and the assessment changes, which is the intended use.

It is not designed to reassure. Its value lies precisely in the fact that it will not.

You enter the hearing having already faced a capable opponent.

The tabbed brief, the compilation, and the prepared opening all remain necessary. What changes is that the question from the bench is one you have already encountered, and already worked through, in the privacy of your own preparation.

That is the entire proposition. The weakness in a case will be found. The only variables are who finds it, and when.

Find yours first.

Devil's Advocate is in private preview.

We are continuing to harden it, and we are opening access to a small group of litigators before a wider release. If you argue for a living, we would value your attempt to break it. Rigorous testing by practitioners is the most useful contribution at this stage.

Available in Early Private Access only.