The Client Who Waited a Decade for a Judgment That Already Existed

The client does not see the institution. She sees the date. The next date, and the one after that. She has rearranged her small business around court appearances for six years. Her children have grown from primary school to college while the title suit sat in the district court. She does not know the phrase 'case backlog.' She knows that her father died before the property was partitioned, and she is beginning to wonder whether she will see the end of it either.

This is not an unusual story. It is the normal story. As of July 2026, more than 5.03 crore cases are pending across Indian courts. The Supreme Court carried 69,766 matters on 1 July. The High Courts, as of 14 July, had 60,62,953. The district and subordinate courts—where the client actually stands—held 4,41,35,357. These are government figures, not estimates. They describe a system in which the average case is not an exception to the queue; it is the queue.

The Queue That Genuinely Isn't Shortening

It would be unfair to say nothing is being done. More courts are being added. As of 30 April 2026, 775 Fast Track Special Courts were operational across 29 States and Union Territories, including 398 exclusive POCSO courts. Another 862 regular Fast Track Courts were functioning across 21 States and UTs as of January 2026. The eCourts Phase III project carries an outlay of ₹7,210 crore, with ₹2,740.86 crore released by April 2026. Over 660 crore pages of court records had been digitised by 31 January 2026. Infrastructure is being built.

And yet no practising advocate should plan on the queue shortening meaningfully in their career. The reason is not a lack of courtrooms. It is a lack of judges. Many fast-track courts have no dedicated judge; regular sessions judges absorb the load on top of their existing dockets. Roughly one-third of sanctioned High Court judge posts lie vacant. District court vacancies tell an even starker story: 29.24 percent in Uttar Pradesh, and as high as 58 percent in Karnataka. India has about 18 judges per million people. You can build a hundred new courtrooms, but if nobody is sitting in them, the date keeps moving forward.

The age of the backlog confirms what the vacancy numbers suggest. As of mid-2026, 71,204 High Court cases had been pending for over 30 years. In the district courts, the figure is 1,01,837 cases older than three decades. More than 10,000 Supreme Court matters have been pending for over 10 years. These are not complex constitutional questions in every instance. Many are second appeals, rent control matters, motor accident claims—cases where the law is settled and the delay is purely administrative.

The People Who Wait Without a Trial

The backlog is usually discussed as a problem of judicial administration. It is also a problem of human liberty. At the end of 2024, undertrials made up 72.6 percent of India's prison population—371,440 of 511,542 prisoners. That proportion was 66.2 percent in 2005. The direction of travel is clear. Nearly half of these undertrials are under 30 years old. About 63 percent did not study beyond class X. Two in three belong to Scheduled Caste, Scheduled Tribe or Other Backward Class communities. These are not people who can afford to wait. They are people for whom the presumption of innocence has become a theoretical proposition, suspended for years while the file moves from one date to the next.

An advocate who has practised on the criminal side for any length of time knows what this looks like on the ground. The client's family borrows money for the surety. The client loses the daily-wage job. The witnesses drift away. The case does not end; it simply ceases to be contested with any vigour. The system records a disposal eventually, but it is not a decision on the merits. It is a decision by attrition.

The Loss Nobody Records

There is another kind of loss that gets even less attention. It is the matter decided not on its merits but on what was found in time. Every litigator has experienced this: the Tuesday night before a Wednesday hearing, going through a precedent file, knowing there is a judgment somewhere that settles the point but unable to locate it before the morning. The hearing happens. The court passes an order. Months later, a colleague mentions the authority that would have changed everything. It was there all along. It simply was not found.

This is not a failure of skill. It is a failure of bandwidth. The Supreme Court and High Courts produce thousands of judgments every year. No single advocate, however diligent, can hold all of them in working memory. The chambers that do better are the ones with larger teams, better libraries, more time. The client with fewer resources gets the outcome that corresponds to the resources deployed, not necessarily the outcome the law requires. The system records a judgment. It does not record the judgment that should have been.

What Actually Changed

For a long time, technology in Indian legal practice meant a laptop with a PDF reader and a subscription to a commercial database. The advocate still had to know what to search for. The database could retrieve; it could not find. That distinction matters. Retrieval is giving you the document whose name you already know. Finding is telling you that a document exists which answers your question, even though you did not know it was there.

The change that has occurred in the last two years is not that machines have become intelligent in any human sense. It is that they have become capable of searching across meaning rather than across keywords. An advocate can now describe a legal situation in ordinary language and receive a set of authorities that address the principle, not just the phrase. This is work that was genuinely impossible for one advocate on a Tuesday night. It required a junior, several hours, and a measure of luck. It is now possible.

The investment figures confirm that this is not a passing enthusiasm. In the first quarter of 2026 alone, legal technology companies raised approximately $2.34 billion across 103 deals globally. By comparison, the entire year of 2025 saw about $2.9 billion, and 2024 saw $1.4 billion. In March 2026, Harvey raised $200 million at an $11 billion valuation, and Legora raised $550 million at $5.55 billion. Capital at this scale does not flow toward a fad. It flows toward a structural shift in how legal work is done. The Indian advocate need not care about Silicon Valley valuations. But she should care that the tools being built with that capital will shape what opposing counsel can find, and therefore what she needs to find in response.

The Verification Question

The scepticism that a senior advocate brings to any new tool is healthy. It should be applied here with particular care. A legal research tool that gives you an answer but cannot show you the judgment it is drawing from is not helping you. It is transferring risk to you and your client. You stand up in court and cite the proposition. The bench asks for the paragraph. If you cannot produce it, you have not only lost the point; you have lost credibility. Credibility, once damaged, does not return in that case.

This is the line that separates a useful instrument from a dangerous one. The tool must show its work. It must link every proposition to a specific judgment, a specific paragraph, a specific sentence. It must make verification trivial, not difficult. If it does not, the advocate is better off with the old methods—slower, but at least honest about what they do not know.

There are tools now, including those being built in India, that understand this requirement. LexPilot, for instance, is designed so that every answer is tethered to a retrievable source. The advocate can read the judgment, not just the summary. That is the minimum standard. Anything less is a research assistant who whispers suggestions but refuses to tell you where they came from. No responsible advocate would tolerate that from a junior. They should not tolerate it from software.

The client waiting in the district court does not know any of this. She does not know what a vacancy rate is, or what a large language model does, or why a particular Tuesday night in her advocate's chamber matters to her case. She knows only that she has been coming to court for years and nothing seems to finish. The advocate who can find the authority that settles her matter—and can show the court exactly where it is written—gives her something the system has largely denied her: a decision on the merits, in her lifetime.