Image [cc] William & Mary Law Library

Niki Black, Jeff Brandt and some others have been having a Twitter dialog on the extent to which computers are currently able to replace lawyer functions. Part of the discussion centers around defining what is meant by replacing lawyers, which falls in to the old debate of what exactly is the practice of law?

Niki offers up a classic criterion as “’advice’ is counsel given to client based on the facts of a case (or legal matter).” At 3 Geeks we have previously discussed some next-generation technologies that I feel fall into this category. However, Niki is looking for simpler, solo/small firm examples. The crux of the discussion is whether there truly is technological solutions to replacing lawyers. I will share my bottom-line thought here: Yes, I think there is. My main point is that the profession has never put much effort into trying to automate lawyer functions.

This discussion brought back to mind the first time I watched a computer practice law. And it was a solo/small firm example.

Back in the 80’s while in graduate school I was a librarian at a branch office of a regional firm. I was in the room when they delivered the IBM XT PC and therefore became the expert on its use (a.k.a. the beginning of my legal tech career).We had purchased the PC as we were a beta site for a document generation system for wills and trusts. A partner at the firm was involved in the start-up efforts of what would become HotDocs. And since this software ran on a PC, we had to get one.

With the program loaded up we began playing with it (which we now call QC). I answered a series of questions about my personal needs related to an estate plan, giving what was essentially ‘the facts of my legal situation.’ Well in to the questioning a yellow screen popped up and ‘gave me advice.’ I do not remember the specific advice, but the gist was that based on the my situation, I should consider changing my answer to the last question about what I thought I would want, since that did not fit with my situation. I recall distinctly sitting back and thinking – Wow. I just witnessed something unique. A computer giving me real legal advice.

So to the Twitter discussion – I repeat my assertion, we haven’t put much effort in to automating lawyers. This tells me there is likely numerous ways in which we can automate. To Niki’s point of view, there is still not much out there. This last point is well-taken. The ability for technology to perform lawyer tasks has been around now for 30 years. Isn’t about time we started using it?

  • Toby: I have a similar experience. In 1988, I bought my first professional PC. It was an IBM 8088 with 128K RAM. I used Borland C compiler; WordPerfect (with rudimentary macro capability) and Ventura Publisher. With this I was able to automate my estate planning practice by writing an expert system, document management and client management systems. It wasn't user friendly–far, far from it–and it really wasn't very smart; but it could create a Will in seconds.

    By 1999 a Texas court in Unauthorized Practice of Law Committee v. Parsons Technology, Inc found that software could practice law. If I correctly recall, the Legislature intervene and crafted an exception.


  • In the ongoing Twitter discussion, Niki is searching for a litigation example. Here is my proposed litigation option:

    I suggest there are 10 to 20 very common claims in a divorce proceeding. Using the same technology from my 80’s example, a series of questions could be developed to determine a client’s fact situation. The result: A draft document where: 1) a portion of the claims will always be included, 2) another portion will be included based on the client's situation, and 3) some claims will be excluded since they don’t meet the client’s situation, or they conflict with other claims the client is making.

    I also suggest this would easily be done using very dated technology.

  • At next week's Ark conference in New York City ('m going to briefly show a system for housing-related litigation practice that began life in the 1980s with the same technology you reference, Toby (CAPS). Quite knowledgeable and intelligent in ways that were already available to lawyers 25 years ago, yet that remain unexploited by most even today.

  • Toby–Thanks so much for drafting this post in response to my request on Twitter for examples of automation for small/solo firms, as opposed to BigLaw. The example that you offer in the post and the one offered in your comment were very helpful.

    On Twitter, I also indicated that examples of automation in the small firm/litigation environment would be helpful, since that was the original point of my blog post ( that instigated this interesting Twitter discussion.

    Keith Lipman offered up this very interesting example in response to my request: (R)isk analysis on settle v(ersus) litigat(ing) in particular subject areas could be developed.

    All in all, this is an intriguing topic. I still stand by my position in my blog post: "Anyone who has followed my writings knows that I tend to agree with this dire prediction. I truly believe that over the next few years, lawyers who embrace and utilize technology effectively–in particular Internet-based and mobile tools–will undoubtedly have an edge over their less tech-savvy colleagues…(However, although) technology can automate and even replace certain aspects of lawyering…at the end of the day, the legal advice and advocacy of a trusted advisor is a decidedly human factor–and, in my estimation, it’s something that simply cannot be replaced."

    That being said, I've enjoyed this discussion very much. Thanks for the conversation!

  • Just as machines replaced humans on assembly lines, I believe that the algorithm will replace much of what professional do. For example, how often do you ask the same questions, explain the same laws, and run through the same flow charts in your head to give legal advice. It happens more often than not. The questions, explanations, and flow charts can all be programmed into a web applications.

    But we'll still need lawyers for now. The web application can probably get the client 80% of the way there. Then the client still wants to meet with the lawyer to check the work and get a warm fuzzy. Hence the big looming inadequacy of products like LegalZoom.

  • I come at this from a Biglaw corporate background, and can't really speak to the solo/small firm examples Niki Black's looking for. In my former life as a big firm corporate lawyer, I used very little automation tech (blacklining software stands out as one of the few exceptions–back in the day, apparently this was apparently done by hand). I think this will change over coming years–a lot of Biglaw junior associate tasks are ripe for automation. They (i) are not very complex and are highly repetitive; (ii) need to be done accurately, and people make mistakes on repetitive tasks; (iii) are high volume; and (iv) are high cost. I've gone into more details on these points here: (

  • By the way, sorry for the pair of earlier deleted comments–trouble getting the link url to show!

  • The sad truth is that more and more machines are replacing humans. I do not think that a computer can replace a lawyer, as this profession requires more than algorithms .
    Yesterday I watched a show presenting some very detailed work, which cannot be done by any machine, however I watched another show about this Asimo robot, who/which reminds me of some sci-fi movie, where robots are taking over the universe. I hope that the time when machines will control everything is pretty far 🙂

  • We risk creating problems when we use terms like "replacing lawyers" and "automating lawyers" in this discussion, because they pit human against machine and create a larger and more dramatic context than we require for this issue.

    I think we should be focusing instead on the identification and performance of lawyer functions. What do lawyers do? Make a list: in any practice area you choose, it'll be extremely lengthy. Then ask: which of these functions can be performed at least in part with an automated process or system? The answer, I think, is: virtually every one of them. And that includes the supposedly safe old standbys like "strategy" and "counsel": huge databases of business and legal intelligence will be able to provide increasingly reliable indicators of how various tactical approaches will pan out. Prediction markets that assess probable outcomes have tremendous potential to change litigation strategies. And just wait till game theory really gets on track in the legal industry.

    Lawyers have never had so many function-performance and strategy-optimization tools as we have today, and we'll have vastly more by 2020. There'll be no point in talking about "lawyer-vs-machine," because there'll be no lawyers without machines. Machines will be so ubiquitous that they'll become less noticeable and less differentiating. The more important question will be: What will be lawyers' value proposition to their clients in an extremely mechanized legal services marketplace? And really, that's the same question we can ask today.

  • I agree with all that Jordan Furlong says and with the article’s suggestion that this subject raises the deeper question of “what exactly is the practice of law?” Practicing law boils down – at all levels of the profession – to assisting clients with various forms of tacit bargaining, i.e., helping clients to take actions and assert positions that will in due course cause someone else (e.g., a potential business partner, judge, or adversary) to eventually respond in a desired way. For example, litigation is, in virtually all cases, simply a form of tacit bargaining (less than 3% of cases ever go to trial – the vast majority are resolved through overt bargaining once the opportunity to engage in tacit bargaining finally comes to an end).

    Several Nobel Prizes have recently been awarded for discoveries relating to what works and doesn't work in various bargaining contexts. This work has led to the development of several new legal-tech systems that are already being used to manage and resolve all sorts of legal conflicts and negotiations at little or no cost and with remarkable efficiency. As Jordan Furlong says, “just wait till game theory really gets on track in the legal industry.“ These systems do not simply generate legal documents or provide advice or recommendations based upon data entered by lawyers or clients. They cut to the very core of what the practice of law is all about.

  • At one point in time lawyers charged for calculating child support for a client. Mow there are child support calculators all over the web. Alimony and support calculators that are based on standards rather than clear rules are already appearing. It is only a matter of time before more complex web advisors become available. see for example Quiz- Should you represent yourself in a divorce? at –