This post was originally published in September, 2015. Sadly, this specific episode of the Lawyerist podcast is no longer available.

If you have any interest at all in the intersection between technology and lawyering, you should really check out this week’s podcast over at Lawyerist (note: recording no longer available), where Lawyerist’s Sam Glover interviews Dennis Kennedy of the Kennedy-Mighell Report. Some of the conversation goes where you’d expect—document review, artificial intelligence, technology versus offshoring, what really constitutes “lawyering”, etc.—but then around the 13:37 mark, the conversation turns to a subject near and dear to my heart; namely, document assembly (which Kennedy apparently has had extensive past experience with).

Although Sam’s not entirely convinced of its value (at one point protesting “I am perfectly capable of automating documents, but in my own practice, I almost never bothered, because it would have only saved me 30 seconds”), Dennis Kennedy responds with what I think are some critical insights:

I think you always have to have volume and … repeatability. And so the best thing is to identify where those things make sense. So I sort of have a couple of principles on document assembly I always think about, that have been part of my learning. I think when you say, ‘What I want to do is find a way to generate the finished document all at once,’ I don’t think you can win with that. So my goal was always to say, ‘Can I use document assembly in a way that generates a really good first draft?’ And by taking the standard to that [level], I think [document assembly] can be really helpful. Although you still have to have volume and repeatability.

Kennedy mentions using Word’s Quick Parts to assemble interrogatories, as an example:

Then I also think there’s this other piece of document assembly that in ancient days people used to call ‘point and shoot’ document assembly, where you’d say, ‘I have this clause file‘ (which, you know, a lot of lawyers think they have or have but they don’t really use) … and whether that’s using, you know, the smart insertion (I’ve lost the name of it off the top of my mind, in Word [Ed. note: Quick Parts]) or whatever where I can go, like, BOOM. I just use a hot key or I click a menu option and I pull those standard clauses in, and I can kind of assemble pieces of documents and other things on the fly to customize them because I do have repeatable components even if I don’t have repeatable documents. So if you said, ‘I’m using document assembly, say, for interrogatories or that sort of thing in a litigation practice,’ you can say, ‘Oh, I see how I can use document assembly for the container of those questions and if I could kind of point and shoot to the questions I want to include, then I’m going to have standard approaches, and that gives me that really good first draft that I can work with.’ And I know I have all the basic stuff, so then instead of saying, ‘Have I included all 30 of these questions or have I remembered everything that should go in here based on my memory,’ it’s all in there and you look through it and you say, ‘Oh, in this case, you know what, we need to ask this, we need to do this,’ and then I think you’re adding the value and turning it from a routine exercise into something that’s actually creative for you and helpful for the client.

At this point, Sam asks a pointed question: “So, why is that we’re still talking about this, though? Like, why isn’t it more widely used and why can’t we say, ‘Okay, it’s just a tool like any other, and yeah, of course lawyers should use it, and they are’?”

Sam and Dennis (starting at the 19:47 mark) take a stab at answering that question in their interview. But what do you think? Why do we see so little document assembly in law firms nearly three decades after that technology’s introduction? Do you agree that many lawyers’ resistance is rooted in a misguided expectation of what document assembly should accomplish (produce a final draft rather than a first one)?

Similar Posts


  1. Hi,
    Re: Lawyers resisiting document assembly.
    A very interesting point!
    Part of the problem is that lawyers like their own style so are reluctant to use someone else’s and even the best prepared precedent sets can contain errors. If a lawyer has prepared a great document, he or she will continue to use it, even if there may be something better or more up-to-date. Lawyers tend to be very resistent to change and the software packages are not cheap. If you are a sole practitioner, using the same documents for possibly a narrow range of work, is a lot cheaper than paying for software, most of which you will never use (rather like MS Office, but that is another discussion). It is also very easy to make mistakes in amending precedents and being an oldie myself, my typing is very slow as I am not trained to do it. It is often far quicker for me to dictate something new each time than amend a previously prepared document, whether it is from a software package or my own precedent bank, although that applies more to reports than deeds. The other problem I have is formatting documents in Word (or rather continuing the formatting someone else has created) so, again, it is quicker and much less frustrating to do it from new.

  2. I purchased Pathagoras 5 years ago and have tried to get disciplined with it in fits and starts but was always too busy to make it work for me, which I really regret. I also find Word Styles to be impenetrable, despite the fact my practice has been paperless for over 10 years and I run a virtual office now with a national practice. So I understand technology. We were doing things in Word Perfect 20 years ago easily that we still can’t do in Word after much effort. Think paragraph numbering.

    So in short, getting good at document assembly takes a fair amount of time and energy, which most busy practitioners are in short supply of. If there were a product that had less of s learning curve and was easy to implement, attorneys would flock to it.

  3. In a previous life I undertook a project to generate all the documentation (including certificates) for the purchase of off-the-shelf companies, appointment of directors, etc. using Ghostfill. The whole procedure took a lawyer approximately 20 minutes (most of which were taken up entering the requisite bits of data) and then (from memory) some 35+ documents were produced automatically. Even armed with what would seem to be convincing evidence of its worth, I never managed to convince the firm to embrace document assembly. However I have to say the junior lawyer involved was extremely grateful and hailed me as a saviour!
    I believe the licensing of the software to be a bit problematic. Ideally you would like everyone to use the software for the portion of their work that lends itself to document assembly – albeit on an occasional and ad hoc basis – but then it’s not cost effective to licence everyone in the firm.

  4. Under pressure (as most are), lawyers have to focus most on winning and serving clients. Until there’s a demonstrable, direct and apparent link between adopting automation tools and bringing in clients, most lawyers won’t invest the time and effort required. In law firms, I think there is a subconscious feeling that the link is tenuous because otherwise everyone would be doing it. We use document assembly as a differentiator, and it is effective precisely because so few lawyers do it well. I work in a small niche law firm, so what we do will never get much prominence so, fortunately, other lawyers will never look at us and say “we need to be doing some of that”. Every time I go to a conference and hear partners of large law firms declare why document assembly wouldn’t work for them, I know we can keep our differentiator a while longer. It’s much better for us if larger firms spend their time wrestling with implementing document management systems, Office 365, etc. Those of us in small law firms need to get our breaks from somewhere! On the other hand, I find this technology is so compelling (in the right scenario) that I find it hard to keep it quiet.

  5. Why isn’t document assembly more widely used?

    From hundreds of conversations with lawyers about document assembly for over 25 years, I’ve learned that the prospect of creating a document assembly system for each repeatable document is daunting. There are any number of useful document assembly engines out there, however, the hours it takes and the review required is demanding. Return on investment can be years away.

    On the other hand “canned” document assembly libraries (such as the DL Libraries which are widely used) can produce sophisticated first drafts for an actual matter after loading the library and completing a 15 to 30 minute Q & A session. And the investment is returned after the first or second use!

    Small law firms and solos can use these libraries to compete with online legal document preparers and bring in related additional business related to the transaction that online prepares can’t offer. And they gain an opportunity to create a long term relationship with the client.

  6. Document assembly isn’t as popular as it should be for several reasons. First, most new documents are drafted in law firms by finding an old one, saving it as a new file name and manually changing it to work for the next case. Even though this approach has a huge margin for error and is slow, it’s habit and habits are hard to break. Second, most document assembly is based upon an interactive interview; and most lawyers have never thought about their documents in terms of what questions must be asked and answered to get exactly the document they want. Our brains can process all of that in seconds, but when faced with reducing it all to writing, it becomes pretty daunting. Third, good document assembly templates are predicated upon understanding and using the high-end functionality of the word processor (like automatic paragraph numbering & spacing, tabs & indents, automatically updating cross references, tables of contents, and tables of authority, etc.). All of that stuff has to work properly every time and I can promise you that only about 1% of legal users have any idea how to control it. Fourth, automating a Revocable Trust (for example) that has hundreds of optional provisions and changeable text is complicated. If the tool being used is sufficiently sophisticated to pull it off, then the tool is also complex and will take a bit of time to learn to use. “I had no idea how much time it would take” or “That program was way more complicated than I expected” are common reasons given for why a document assembly project was never finished. Fifth, as Mr. Glover points out, what incentive do I have to automate tasks if it reduces what I can bill a client? That dilemma is commonly known as the Productivity Paradox. The solution is to stop believing that how long it takes you to do something has any relationship whatsoever with the value of what you’re delivering to the client. If a task takes an hour and you charge $250/hour, then $250/hour is your effective hourly rate. If you charge a flat fee of $250 for a task that at used to take an hour but now takes 10 minutes thanks to automation, then your effective hourly rate becomes $1,500/hour. How’s that for inspiration? Sixth, many lawyers feel that dictation is more efficient because they can talk a lot faster than they can type. However, this fails to take into account the extremely high costs associated with dictation/transcription. Payroll is every law firm’s number one expense. If the process I’m following involves other people (transcription), then the costs are high – particularly when the transcriptionist isn’t billing for his/her time. Lawyers dictate, it is transcribed (the average hourly cost for a support staff person including taxes, benefits and overhead is $27/hour) and printed (costs for paper & toner/ink). The lawyer then marks up the printed copy and the changes are entered into the word processor by the support staff person (non-billable) and the document is printed yet again. Sometimes, a 3rd round of changes are entered and printed. By contrast, a document assembly system can generate documents in 1/10th the time and cost, I don’t care how fast someone dictates. Even if voice recognition is used to transcribe the dictation, it still has to be corrected and formatted. Both of those tasks are avoided with document assembly. As an aside, I strenuously disagree that document assembly is only good for a first draft – I’ve seen hundreds of systems that produce final drafts the first time with no post-assembly editing required.
    Ultimately, the end-game is totally worth the automation investment if you have the volume to justify it. There are obviously a lot of practice areas where the deliverable IS a set of documents. If you can dramatically increase speed and accuracy, reduce turn-around time, and improve profitability, then it’s worth overcoming the six foregoing stumbling blocks.

  7. Most small law firms do not have the repeatability of documents to warrant creating their own documents with a document assembly software engine. My grandfather, who practiced law in a storefront for 40 years in the mid-20th century, charged $50 to do a will. When work for the client was finished, the bill was usually between $200 and $400 because he uncovered additional needs of the client (estate planning, powers of attorney, property transfers, etc.). Today, exposure to this type of client is limited because the mass market advertising of online preparers draws clients away from lawyers.

    Document assembly can play a role in bringing these clients back. You need a program with the familiar language lawyers’ use in transactional documents, the
    Q & A built in; and one that produces an initial draft after the few minutes it takes to enter the answers to the questions in a questionnaire sent to the client prior to the interview, or entered in a session with the client.

    Because there is no investment in time, only the cost of the program, you can build volume as you attract clients with a low fixed fee and recover your investment after two or three clients.

  8. First, nobody is doing it to save their clients money. They are doing it to make more money. This job is not a charity.

    As for why it rarely makes money, the reason is circular:
    1) Setting up document assembly is difficult and time consuming. Mostly time consuming.

    2) If you make a high hourly rate and have a good amount of work, you generally don’t want to spend the (large) amount of time needed to set it up, define clauses, etc. I could do it myself, but there are rapidly diminishing returns for me to do so: Having ten quickparts and a few autocompletes is fast and efficient and will save me 50% of such issues time for 10 minutes of work. Setting up all my commonly used documents will save an additional 30% and will take ten hours.

    3) So the job gets done by low end attorneys, or non-attorneys, with minimal supervision.

    4) As a result, the people who would tend to use it (high priced attorneys who can add efficiency) won’t do so because they don’t trust it.

    I’m a computer geek and I could easily adopt pathagoras or whatnot. I don’t have time to set it up, though. And if I give the task to my paralegal or associate, I won’t trust the result. So I don’t use it at all.

  9. Thanks for all your good work… I found this section extremely interesting and filled with excellent descriptions of how to work effectively with technology.

  10. As mentioned above, document assembly programs take a long time to set up and are only good for documents that have a high percentage of ‘boilerplate’ text. As IT, we see the value of not copying old documents (and their inherent corruption), but most attorney’s don’t have to solve the problems that creates, so it’s not a good marketing point. When I compare an expensive document assembly program + the development time + the training time + the marketing effort required for wide adoption to simply offering the boilerplate text as a clean read-only template + publishing a building blocks document with a clause library so the lawyer can use Quick Parts/Autotext to insert the variables (which is all free and takes about 5 minutes to learn), it’s no contest; I’ll use the free and easy tools every time!

  11. Where I practice many, if not most, do use document assembly. They just don’t create their own. They use programs specifically designed for their practice area. For example, most family lawyers here use Form Builder by West. Some use TexDocs. Many (but a much smaller percentage) estate planning attorneys use Form Builder or one of several other dedicated estate planning document assembly programs.

    I know of very few who create their documents using the Word features.

    The biggest problem that I see is people not treating the output as a draft. They just clean it up a bit and send it out. The result is documents that aren’t as relevant as they could be. For example, I frequently see “canned” interrogatories or requests for production that don’t match the facts of the case.

    The problem with doing it yourself (as opposed to using a program with the forms completed) is that the forms continually change. The provider that we use provides thousands of forms and each time the legislature adjurns they update hundreds of them. There is no way that I could do that. I would bet that I only regularly use about three dozen of the forms that I subscribe to, but I also use many of them on an occasional basis. This is far to many to invest the time to create the assembly system myself.

    I, for one, would not like having to go back to the days of creating documents individually or using prior documents and changing them for each case. This does not even take into account the vast opportunity for mistakes when you cut and paste language.

    Thanks for the thoughtful analysis. And thanks to the others who respondent with cogent comments.

Comments are closed.