The best legal technology is the kind you stop noticing: Nyayassist

Legal AI has arrived loudly, and it announces itself in every interaction, doing the opposite of disappearing.
Yash Rane, Gaurav Nagotanekar
Yash Rane, Gaurav Nagotanekar
Published on
5 min read
Listen to this article

Every law firm has a version of the same person. At one firm we know, he has been there for twenty-six years. He does not attend strategy meetings. But when a partner walks out of a hearing and says the name of a matter, the file is on the desk before the partner has finished taking off his coat. Not the file cabinet. The file. Open to the right page.

Nobody at that firm calls this a system. It has no dashboard. And that, precisely, is the point.

In 1991, a computer scientist at Xerox PARC named Mark Weiser opened an essay with a line that has quietly haunted technologists ever since. The most profound technologies, he wrote, are the ones that disappear. They weave themselves into ordinary life until they become indistinguishable from it. Weiser was thinking about computing, but any lawyer would recognise what he meant, because the profession is full of examples.

Consider the photocopier. It arrived in Indian law firms without a launch event and rearranged the practice of law more thoroughly than most software has managed since. Discovery changed. Briefs got longer. Entire habits of preparation were rebuilt around the assumption that a document could be in two places at once. And yet no advocate has ever stood before a judge and said, "My Lord, I photocopied this." The machine did its work and then vanished into the background of the profession, which is the highest compliment a technology can be paid.

Now consider how legal AI has behaved. It has arrived loudly, and it announces itself in every interaction. It asks the lawyer to leave the place where the work actually happens, open a new tab, learn a new set of habits, phrase a request the way the machine prefers, and then carry the output back by hand. It is a destination rather than an instrument. It has, in short, done the opposite of disappearing.

We say this as people who build the stuff.

The evidence is not hard to find. Walk into most firms that bought a legal AI licence eighteen months ago and ask, quietly, how many people still use it. The answer is usually a number smaller than the seats they paid for. The industry has a polite phrase for this. It calls it an adoption problem, which neatly implies the fault lies with the lawyers who did not adopt.

We think that reading is exactly backwards. A tool that has to be adopted is a tool that has not fitted. Nobody ran an adoption programme for the photocopier.

Here is what we have come to believe after building for the Indian bar, and after enough hard conversations with senior partners to have lost any romance about it. The unit of legal work is not the prompt. It is the workflow. And workflows are not generic.

Two firms of identical size, in the same city, practising the same area of law, will run a due diligence exercise in ways that share perhaps sixty per cent of their DNA. The rest is the accumulated judgment of a particular set of people over a particular set of years. Which flags matter. Which is noise. Who reviews what before it moves. How a red flag becomes a note, and a note becomes a call to the client. That residue is not inefficiency waiting to be optimised away. It is the firm. It is what a client pays for when they choose one firm over another.

Standardised software has to ignore that residue in order to function. It has to assume all firms are one firm. Which is why, when the demo ends and the real matter begins, the tool feels like it is asking the practice to bend around it. And a practice with a hearing on Monday will not bend. It will go back to the way it did things before, and the licence will sit there until renewal.

Technology that disappears has to be shaped to the place it disappears into.

This is the thinking behind NyayAssist Enterprise, which we are opening to firms now. The core of it is not a longer feature list. It is custom workflow automation. We map how a firm actually works, in its own vocabulary and its own sequence, and we build to that shape rather than asking the firm to conform to ours. Around it sit the things a firm cannot practise without: collaboration for teams working the same matter, so the work moves between people the way it already does on paper; purpose-built diligence workflows; and a security posture appropriate to institutions that hold other people's most sensitive material as a matter of professional duty.

That last point deserves more than a clause. In our experience, the question that ends most enterprise conversations early is not "what can your AI do." It is "where does our data go." Confidentiality is not a feature of legal practice. It is the condition of it.

But everything we have written so far is a claim, and claims are cheap. The nature of our argument makes this worse, not better. Software that disappears into a firm's work cannot be shown in a forty-five minute call. You can only tell whether it disappeared after it has had somewhere to disappear into.

So we have built the offer to match the argument. Firms can run a thirty day pilot of NyayAssist Enterprise at no cost and with no commitment. During that period, we map the firm together and work towards building what we find. In the end, if the firm does not feel the difference in its own work, it walks away owing us nothing. Only if it does work do we talk about scaling. The risk sits with us, which is where it belongs, because we are the ones making the claim.

Which brings us back to the man with the file. What he offers is not speed, exactly. It is the absence of friction at the moment friction would cost the most. The partner does not think about him. The partner thinks about the matter. That is the whole of it.

If, in five years, the lawyers using our software have stopped talking about our software, and are talking instead about the case, the client, and the argument they are about to make, we will have done the work properly.

Until then, we are just another tab.

Yash Rane and Gaurav Nagotanekar are co-founders of NyayAssist, a legal AI platform based in Mumbai.

Legal teams can book a thirty day NyayAssist Enterprise pilot, at no cost and no commitment, at nyayassist.ai/enterprise-legal-solutions. We map your firm's workflows and work with you to build NyayAssist around the way you already operate.

This is a sponsored post.

Bar and Bench - Indian Legal news
www.barandbench.com