[pnmahoney]

Really?? The move away from standard billing had always felt like it would "always be in the future"; but I take that back if that's now A Thing in lit heavy on case research. Still assuming that ^ research is better facilitated by pagerank+semantic centrality, as opposed to the boolean features that the attorneys who do the most motion drafting work are already most well-versed in. The most valuable usage I'd heard of this type of product (the adoption of which I'd had reasons to keep asking extensively around about) was from my pal in international arbitration whose colleague wanted to do a search across courts in one screen. So maybe that was it. ¯\_(ツ)_/¯

[rayiner]

Yes. Even in the context of billable-hour matters, it is typical for firms to write off hours that exceed certain targets per sub-unit of the matter. That is, in effect, a shift towards pricing based on unit of work (e.g. a motion to dismiss) rather than pure hours.

I disagree about pagerank being appropriate for legal research. Pagerank-type algorithms will help you find the seminal, widely-cited cases in an area. But if you want background law, it is easier to just consult a relevant treatise or other reference book that has those pre-digisted and laid out for you. What really takes time when doing legal research is finding cases that have features in common with yours: similar procedural posture, similar factual wrinkles, etc. Usually, you're looking for a way to argue the opposite of the general rule set forth in the seminal cases. Pagerank-type algorithms won't help you find that; boolean searches on specific phrases will.

[pnmahoney]

I'll own that I could see why you might think that I think a certain way but consider that we do not disagree? I pointed out that - with a premium on billing efficiency - the linchpin the product's actually delivering value in such an environment would still rely on (what I note as) _the assumption that_ a semantic+graph -based product does in fact cut down research hours, compared to the boolean-based approach.

Your explanation about the value of exceptions to rules in research is correct and would occur to, say, pretty much anyone who has drafted a motion (worth clarifying: not a drag on you). Far from a disagreement, the truth of this premise is why - as I pointed out - the most experienced/valued appellate drafters effectively rely on boolean. (Except, I guess, the 'within' operator might count as a semantic operator. Although that's been in lexis and west for as long as I've seen them, and I suspect have found immense use for much longer than that...)

edit: Adding that of course only one of the two following premises need be incorrect (firms should use products which enable them to spend less time on research; the citation graph enables lawyers to spend less time on research while yielding identical quality; firms should use citation graph -based products). Depending on your circumstances, only one may in fact be wrong.

[vqc]

Law firms are now internally tracking the number of hours spent on units of work (e.g. motion to dismiss, MSJ, etc) so that they can provide much more accurate pitches and estimates of cost.

On the latter point: https://www.judicata.com/ understands legal research and has built a legal research tool in the right way.