Legaltech landscape

General Systems Unit™ →Go uptown A quixotic attempt to change the world, one iteration at a timeIndex: Click ᐅ to expand: General Systems Unit Design Tell me more Sign up for our newsletter — or just get in touch: for ½ a weekly 🍺 you get to consult JC. Ask about it here.

With a hat-tip to Radiant Law’s Alex Hamilton for this categorisation of the contract process in his excellent book Sign Here, here is a functional breakdown of the contract process as Alex describes it, mapped against the contract tech landscape, as we find it.

As Alex points out, any of these functions can be captured by more than one application — which is itself a commercial problem for legaltechbros, because no-one likes to pay eye-watering rent annuities for products which are partly duplicative.

By way of prediction as to what will fly, what won’t, and what will lumber along the ground like a rhinoceros flapping its miniature gossamer wings and wondering why she cannot get airborne, the JC has added commentary along the lines of cui bono: who benefits? Management or Eagle Squad? — how hard is each to implement, how readily will a legal eagle take to it, if it is implemented, and to what extent does its careless implementation aggravate problems the installation was meant to solve?

There is an argument that the trick the legaltechbros are presently missing is consolidation: there are (literally; trust me) hundreds of startup variations with differing combinations of template management, document assembly, document automation, contract review, contract approval and digital execution — you would expect this, as they are contiguous steps on the same commercial process — and all of them leave something to be desired. The entrepreneur that can rally these vendors together and consolidate them into a coherent single product — along the way crushing the precious aspirations of so many entrepreneurial plodders who dared to dream — might have a proposition on its hands. Especially one who builds it on an open-source platform. Like that will ever happen.

The colossal collision of interests that needs to be resolved in order to deliver true front-to-back processing brings to mind Stuart Kauffman’s concept of the “adjacent possible” — being possible worlds that are accessible directly by opening one of the doors presently available to you — a remark on the path-dependency of any evolutionary change. From most organisations, integrated front-to-back processing as a concept is not remotely adjacent to where they are now, dealing with a hodge-podge of locally implemented databases, applications, systems, processes, workflows, hacks and work-arounds — ranging from fully weaponised SAP or Tableau implementations to spreadsheets with jury-rigged macros stored on the C: drive, each developed by different teams and individuals, without thought for the wider process, to suit local and probably temporary needs, with a view to putting something better in place when budget allows.

Bear in mind the business units that develop these hacks themselves morph, merge and wink out of existence, as a function of the tectonic political machinations of the organisation, as do regulatory and commercial imperatives that notionally prompt them thus the system changes itself over time without any particular rhyme or reason — this is a truly dynamic, dancing landscape — and you can see that system is not only certain to be a monster, but it will categorically get more monstrous over time.

A one-stroke attempt to fix even a component of the system as a time-bound project can only have a temporary effect, and even that will be offset by the distraction value and implementation glitches associated with that project.

The possibility of even building the consensus to implement a comprehensive front-to-back system, let alone specifying it, let alone actually implementing it, and more to the point, persuading all the squabbling stakeholders to abandon their present system work-arounds, would be quite impossible even if your desired end-game was not a moving target. The localised, ramshackle, rag-tag fugitive fleet of hacks and workarounds has that advantage: it can and generally will flow, meld and morph, however inefficiently and unprettily, to cope with whatever demand is thrown at it. It has at least the flexibility of a bottom up ecosystem: top-down central planning will work hard to beat that in the long run.

So, answers?

One is to invite all stakeholders to imagine a perfect, machine-age end state, however fantastical, and encourage them to develop their own systems toward that. You may never get to that destination — it may be impossible, or have winked out of existence, or lost its allure by the time you get there, but at least open the nearest door in that direction that puts you in a better place than the one you are in now. Evolve away from an unsatisfactory now: nudge yourself in the direction of what looks like a more satisfactory future.

That may change your options, but it doesn’t cut off them off — every room in the palace of adjacent possibilities has many doors opening off it.

This is a lesson both of the open architecture design of the internet (“end-to-end principle”) and also of the famous Bezos memo.

So, cui bono, Sonny?

Phase	Function	Description	Management appreciation factor	Implementation hassle	Lawyer acceptance factor	Iatrogenic factor
Phase of contract process	What is the innovation?	What does the bit of kit do?	How excited will management be about this?	How much of a pain in the fundament is getting the kit in, setting it up and getting it to work?	Once implemented, how realistic is wholehearted embrace?	Risk of inadvertently entrenching rent-seeking behaviour?
Initiation	Legal fees bid management system	An auction portal for seeking competitive bids on external legal advisory projects	High: delivers control, visibility, metrics and cost pressure on external counsel	Medium: This all falls on legal operations though, so you may confidently assume it will be done.	Low: While no work to implement for lawyers, it removes autonomy, power to choose counsel, is unpopular with outside counsel as any bidding system guarantees more losers than winners.	High: Asks wrong question, namely: “how do I get the cheapest legal advice?” rather than “how do I get the best advice?” or, for that matter, “do I need legal advice at all?” But will it work, in the short term, in bringing down legal spend? Yes. Will it create an alternative structure that is arguably worse? Yes.
	Self-service portal	A place where Sales can go to get pre-appoved forms to send out without a legal eagle having to vet them first.	High: Speeds things up and pushes away low-value work from Eagle Squad	Medium: In theory, easy: could be as easy as an intranet page with a drop-down and a “submit” button. In practice, hard, as you need to find someone in the IT to do it, and making any change to the tech stack requires a scrum, a sprint, six months in a sandbox, 36 person hours and a budget of $300,000. We will get onto that when budget allows.	High: anything that pushes NDAs off the desk has to be a good thing.	Low: Disintermediates unnecessary legal touch.
	Initiation and prioritisation tool	A queueing app for prioritising prospects, gathering information and setting terms before negotiation starts.	High: Good oversight of process, good MIS	High: It’s a tech and operational job, but it will span several silos with different owners and conflicting priorities. Design, ownership of and responsibility will therefore be an unholy fight. Ultimately likely to be over-specified, requiring inputs for management gratification that are not on the critical path, so likely to fall quickly into disuse, if it ever gets off the ground.	Medium: If it works — big if — it will head off an inordinate amount of pre-negotiation faffing, chasing people up for input, and sets a clear business priority.	Low: Disintermediates. Provides simple information inputs and disciplines personnel to follow process. Likely outcome: 18 months of stakeholder distraction, followed by discreet climb-down and no further mention in dispatches.
Drafting	Template Management	Centralised templates database or clause library for approving and warehousing approved forms and boilerplate.	Medium: Should drive efficiency, but part of that ineffable world of legal eagles that management doesn’t understand.	High: Quite a lot of implementation, and legal ops will be disinterested. Once implemented, a lot of work to sort and upload templates, compare them, weed them out, and assign owners to them, which no-one will want to do.	Medium: A useful productivity tool once implemented, saving a lot of tedious bureaucracy, and a good platform for standardisation and quality control later on.	Depends on how implemented: this could be a fulltime career for a squadron of nosey parkers, or it could be a light touch self-service tool.
	Document automation	A preconfigured questionnaire to generate first drafts of standardised contracts.	High: This is genuine high-five, look-at-me stuff: potential for handsome MIS is great. Also, it pitches really well. Legaltechbros love this sort of thing.	Very high: requires input from lawyers, legal ops and legaltechbros. You have to extract the logic from your templates, code it, and build a machine to make it. Ongoing maintenance a chore, too.	Theoretically high, practically low: one of those things that seems great in concept, but sucks in practice. User becomes a form filler-outer: no legal eagle wants that.	High: highly likely to take a bad contract situation and make it worse. Maintenance of templates becomes an IT ticket: expensive, slow, and out of lawyers’s hands.
	Freehand Document Assembly	An add-in to word to allow lawyers to access the inhouse clause library to quickly assemble novel drafts from standard building blocks.	Low: What do I care? All the cool stuff is a function of the template management system it feeds off (if you haven’t got one of those, forget about it).	Low: Assuming you can find the software, it is a straightforward plugin. This is the basic promise of a distributed end-to-end system.	High: This is a neat tool that saves time and ensures I don’t forget anything. What is not to like?	Low: Disintermediates nicely.
Negotiation	Document mananagement system	A matter management system for creating drafts, storing emails, documents, version controls, and collaborating.	Yuuuuge: This is the daddy. A DMS promises to give the lawyers infinite productivity and workflow, while delivering management total detail about what every lawyer does.	Immense. A multi-year project to extract your legal team from infrastructure the rest of the firm uses and put them on a “better” system. And that’s before you try to integrate it into your external spend control regime.	High in theory, low in practice: Legal eagles think this is what they want: when presented with the finished article, they find their byzantine folder structure in Outlook PSTs wasn’t so bad after all.	Higher than it ought to be. Suddenly there are information security officers, usage monitoring metrics, champion groups and stakeholder surveys as far as the eye can see.
	Automated contract review	AI — call it neural networks, machine learning, or some school-leavers from Bucharest with an xml schema — reviewing and marking up standard form contacts.	High: Because it presents, misleadingly, as low-hanging fruit, legal operations folk glom onto it as a way of making waves. Plus, the GC hears “AI” and thinks “HAL 9000” and not “glamourised deltaview plus temps in Gdansk”, as she really should.	Low: If you make the mistake of displaying any interest, replying to their email, vendors will be at you like a plague of locusts and will never let you go. All you need is a playbook and to remember the email address.	Low: It reduces legal eagles to form-fillers, the form takes too long to come back, and it’s easy just to do it yourself. Full analysis here.	High: A job your team used to do off the side of the desk now costs £400k annually, and requires a weaponised procurement and internal audit system.
	Manual review assistance	Comparison, formatting tools,	Low to nil: This is not exciting, the procurement people will hate it, and IT will say things like “why don’t you just use document comparison function in word?”	Low: Usually a plugin to Word, though undoubtedly will be some configuration clashes with some other filters and metadata analysers already in use.	Medium: lawyers ought to love the productivity bump: they’re good with deltaview, but tend to leavev the formatting and janitorial stuff alone.	Low: No humans required. Genuine disintermediation. Puts power in lawyers’ hands.
	Onboarding portal	Front-to-back onboarding portal for tracking the progress and status of onboarding, with milestones, logging all escalations and recording legal reference data.	Supremely: this is the holy grail of legal operations: every impulse reduced to a fully trackable and auditable series of numerical data points, suitable for RAG status indication by PowerPoint.	Enormous: it will involve reimagining the interaction of seven departments, each wedded to its own incommensurable processes built on, variously, email forms, spreadsheets, playbooks, the command prompt and a Wang mainframe that was due for decommissioning in 1983 but survived and is now so integral to the firm’s systems than it has its own round-the-clock armed guard.	Low: Reduces lawyers to disempowered form-filler-outers and button pushers.	Medium: Imposes a mammoth superstructure over something that is already too elaborate, and ducks the question why your credit officers are embedded in a manual process where all they do is push a button. Rather than institutionalising button-pushing could you not rebase your terms?
Execution	Execution approval tool	Online tool for gathering approvals for execution of documents at completion of negotiation.	Medium: Really just an unambitious attempt at an onboarding portal.	Medium: But mainly because you’ve done a half baked job. You claimed to have done an entire cross-silo process reimagining, but you haven’t really, have you.	Medium: Relatively low, perhaps, in that while it will involve a lot of work collating every escalation made over the course of the negotiation and putting each one into a digestible approval form, this is nothing really that negotiators aren’t obliged to do in a fully manual process. But the potential change in outcome is not nearly big enough to be worth the incremental effort.	Medium: Not an enormous additional burden, but it doesn’t get rid of a lot of of encrusted iatrogenesis in the process already.
	Digital execution	Creating a means of digitally executing legal docs to capture who signed, when, and under what authority, thereby capturing text-encoded final docs (as opposed to scans).	Low: Again, it really should be integrated into the approval and onboarding workflow tool. By itself, it is fiddly and doesn’t offer a great deal of bang for buck.	Medium to high: Depending on how thorough you want to integrate the digital execution tool into the firm’s personnel and authorisation database.	High: Once you get over a the small comprehension hurdle, this will be popular. Saves all that tedious faffing around with printers, scanners, incumbency certificates and filing and so on.	High: Digital vendors price their services at an outrageous per-usage costs.
Contract Management	Contract and metadata management	Golden source storage system for executed contracts, ideally containing detailed catalogue of material legal and economic terms (this required “metadata extraction” below)	Medium: To a bean-counter who has not apprehended the horror of deciphering a heavily-negotiated legal agreement, a comprehensive metadata project seems intuitively brilliant. This should really be part of a onboarding workflow tool, though that is fantastically hard to achieve.	High: Depending on how you do it, this could be a five-year, $30m project occupying an entire army of school-leaver consultants in Albania, or a metadata-extraction projection by some preternaturally useless machine learning application for which you will be charged, handsomely, by the document. Neither will deliver in a usable product.	Low: Legal eagles have no interest in structured metadata, which they regard as a form of modernist cheating. Same goes for the senior executives in risk management functions who ask for this kind of data. If you have to understand what is in a legal contract after it is inked — God willing, you never will — there is only one way to understand it and that it by reading it and then helpfully summarising it in a table, in Microsoft Word.	Colossal: This is what rent-seeking was brought onto God’s green Earth for.
	Obligation management	Taking the material terms of your legal contracts, that you codified in the metadata management phase, and feeding them into your risk and capital engines.	Medium: A nice touch, though most risk management functions independently monitor their own risks using existing systems. While of course they should key off a golden source, that assumes you have a golden source and it is reliable. Few live risk metrics key directly off executed legal contracts (as opposed to being reflected by them). Close-out netting is a notable exception.	Low: Generally limited to close-out netting, and that will generally already exist, given how important it is. But the more information that is designed to feed from the legal database, and the more systems it feeds into, the harder the job.	High: Saves a tedious job, and triages stupid questions. Happy days.	Low: Monitoring will mean someone in risk needs to understand what is going on, but assuming you limit yourself to a small number of key pieces of data, okay.
Tracking	Legal term benchmarking	Heatmaps for determining which of your contract provisions are most hotly negotiated	High: Plays to the “high-modernist, I can control everything by data” mindset.	High: involves you having implemented a ton of other innovations first. If you have, and they are working, it should be easy enough to generate. But you won’t have, and they won’t be.	Medium: if lawyers are allowed to freely interrogate the data. But, being “data” it will be kept at arms’s length, owned by someone in IT, and they will have to raise a ticket and wait to days two see any of it, by which time attention will long since have wandered.	High: In order for it to be available, let alone working, there must already be a military-industrial complex of rent-seeking already in place.

See also

• The Bezos memo
• Adjacent possible
• iteration
• The temporary tends to become permanent
• When budget allows
• Why is legaltech so disappointing?