Technologists and (Lawyers) Do Not Understand About Legal Services

Document Automation and Client Expectations

With Ailira we were able to generate automated documents efficiently and cheaply for clients. They could ask questions, tailor each document themselves, read other information and get all the things they wanted provided necessary for completing what I thought was the legal service: the production of the document. This could be achieved online.

However, only a very small proportion of people wanted to complete the document online when given the choice. Instead, they wanted to come into the Law Firm lawyers in one of our various locations, or after the COVID19 pandemic more commonly many accepted video calls. I came to the realisation that the client wanted was not a piece of paper, the document, that I thought that they were seeking. Sure, they were paying for a legal document. But what they really wanted was something else. The client wanted piece of mind, that their legal needs have been met.

For a Will, the person wanted peace of mind that their affairs would be looked after upon their death. They wanted to see a person so that they could eyeball them and get confidence that their affairs were being looked after. They need to understand that the system was created by lawyers and was legally effective and therefore would have that correct legal nature. But ultimately, having a piece of paper with lots of words on it is not something that many people aspire to. Their needs are in a general social construct. This is something that is misunderstood by technologist that make wild prediction about law being automated and also misunderstood by lawyers who produce this services (although most lawyers intuitively understand what they are providing).

The Myth of Legal Automation

They are constantly comments that law is going to be automated. ChatGPT will replace all lawyers. The premise being that ChatGPT or other LLMs can generate lots of words and law is constituted by lots of words, so therefore these words can be automated. Not so.

See the task of a lawyer is rather incidental to the task of generating words. The classic marketing example is a drill company does not sell drills. No costumer has need of a drill. Instead, customers have needs for holes in things and drills are the mechanism for which they meet that need. Customers do not buy drills instead they buy methods for which to produce holes in things. Lawyers do not produce pieces of paper. Lawyers deal with legal relationships between people to safeguard their legal health, in a way a medical doctor might safeguard a persons physical health. Lawyers resolve existing disputes and attempt to prevent future disputes.

So long as there’s discord among humans then there will be the need of lawyers. Dispute is not something that can be automated away because it is ultimately and emotional and relational matter. The tools which I used in the dispute though can be automated. But the result is not to decrease the amount of time or energy in the dispute.

The Ever-Increasing Contract

If you review contracts from years ago, ie before typewriters, their brevity is readily apparent. Each sentence in the contract must be handwritten in which case it is laborious to expand on the contract. The initial terms need to be written by the author or dictated by the author. They began to expand with typewritten and pro forma documents. The ability to preprint a template, reduced the cost per word, but had the effect of increasing the number of words created. The clearest example of this is in the greatest technological innovation of the last 50 years, which is being the ability of copy, paste and delete.

Prior to the 1970s and 1980s all documents would be dictated and typed and if there was an error the typist must retype the document, or at least white out some small part. New parts of documents might be dictated or described to be retyped, or photo stated as a template, but large scale customisation was difficult. However with the ability to copy, and paste, and delete, the size of contracts vastly increased. The cost of adding in an interesting clause or customising a clause was negligible. The template can then easily be reused and recycled and contracts become ever more voluminous, subject only to attempt to rework them for readability particularly in consuming facing environments.

The same is applied in litigation. If you read judgement from 200 years ago they are short and terse. Major hearings for seminar cases will be half a day. Now judgements go on an on and litigants might be in court for months as they real out hundreds of thousands or millions of documents that have been created easily though copy and paste and found through various methods of search and collection.

The Right Expense for the Right Product

When I was a young lawyer I was told by (someone cynical) Counsel that disputes would continue until the parties are exhausted, and then they would settled. The amount of exhaustion that a party could withstand, would depend on their overall resources, and the wealthier they are the greater expenditure they could withstand.

This means that for very well wealthy parties they would spend some certain amount on litigation, therefore for a particular dispute automation would not change that. The more automation there is, the more pieces of paper might be produced, there would just keep fighting until they are so exhausted. An example is estate disputes. The fundamentals of an estate dispute is that one or more parties do not feel aggrieved that they did not get the right amount of inheritance befitting their relationships and needs. The greatest predictor of how smoothly an estate goes, is the emotional states of the parties beforehand.  That is regardless of wealth if a party is happy in their relationships with their family before somebody passes then they are quite unlikely to dispute the estate. Conversely if there is discord in the family then they are likely to have some dispute. They will dispute to the amount that they can, expending such resources that they can until they are exhausted. The larger the estate the larger the amount of exhaustion that the parties will seek.

This creation of a dispute is one of the reasons why many jurisdictions have created legislative avenues for disputing inheritances. Prior to having a mandated avenue the only method of disputing an estate would be if there is some technical defect in the will. Accordingly, an unhappy party would applaud closely over the word of the will in attempt to find some defecting and get that will throw out. They will then attempt to sue their lawyer or the lawyer of the decease or the disadvantage party on attempt to sue the lawyer of the deceased for drawing a will that has been successfully challenge and it’s negligence.

Instead of having parties force to attempt to find fault in the work of lawyers the ability to make a claim on an estate in the event of inadequate provision provides an avenue for parties to vent their anger and rage. Legal automation might make it easier and cheaper to review termism of wills and compare them against inheritance and family provision legislation and cases and suggest where parties are aggrieve and how much they might get and discover documents and prior statements of the deceased and collate them into Affidavits and reduce the cost of the work.

However, the likely output of this is to crate an opportunity for greater quantities of product and have more expensive disputes.

An Elastic Demand

What we should not confuse is the movement of demand and supply curves. If legal automation reduces the cost of production of end legal service (separate to the product, being the resolution of a dispute past or future) then we are simply moving along a demand curve that it is unchanged. Given the demand curves are down with slopping, that it is to say, at a lower price there would be a greater quantity demanded, then we know that if legal services are cheaper to produce then there will be more demanded.

Contract Review

Another example of the demand for legal services being dependent on the use case and not the output is in review of contracts. The contracts are automated and assigned daily with no review. No doubt there is an update to terms of services that you have signed with almost no review, even after clicking (falsely) that “I have read and understood”.

The reason why there is no review it is because the products that they relate to are of low value. A software subscription for a few hundred dollars does not really matter if it has an honourous dispute, clause, because in the event that you have a dispute it is not worthwhile for the sake a couple of hundred dollars. And in any event the software, we will simple cease to use the software and find some alternative. But the higher the value of the contract the greater the scope of review. Imagine that you had one contract that would determine your income over the next 20 years, say for in a relation to employment or business opportunity. Weather this contract is agreeable to you and suitable will vastly impact your lifetime earning and success.

Any party will spend as much resources as they can master to have such a contract reviewed carefully. Even someone who is impoverish will seek some kind of expert advice, even if the greatest expert that they can master is legal aid centre or a wise friend.

For a successful business, they will spend thousands of dollars upon a lawyer to review and negotiate such a contract. For a multinational firm the price to review such a contract which will impact in the success of the firm over the next decade or two might be in the millions. Even if the same template contract was presented in each case there would be an extensive review and the level of review between someone who is impoverish, a small business, and a multinational business would scale, even though the work to be done might be theoretically the same.

If ChatGPT could review and advise on the contract, a multinational company would not accept it they would instead spend of millions of dollars on consultants to review and negotiate a contract. Indeed with the benefit of technology, they would expect that their advisers would be using automation to find other contracts and other things to review as part of the negotiations.

We can think of this as a form of insurance. What the legal services are providing is insurance against some kind of downside risk. In the case of transactional work against future downside risks, and in the case of litigation attempting to minimise present downside risks. There was an amount that parties are willing to pay to reduce that downside risk. The greater the downside risk the greater the amount that they would pay.

If the parties are in litigation the dispute arises out of some emotional or relational issue that they have producing the downside risk given their particular circumstances. Most people do not have disputes and spend their life in court. It is a minority of people who have a legal dispute and could potentially have a legal dispute. The fact there is a dispute comes down to a mixture of their personality, their personal relationships and their relationships with the outside world. This is not a criticism, as indeed it is important that there is tension, and people takes risks when dealing with potential bad actors which results in a risk of litigation.

However, given those sets of circumstances there is an amount that is a cost that must be borne from those circumstances and should the party wish to maintain their legal position and relationships then they must pay an appropriate legal cost. Automation will just increase the amount of legal services that they can utilise although it won’t change the ultimate amount that they spend.

Non-Discretionary Services

Where there can be savings through automation is where there is a largely non discretionary purchase. For example, where there are expenses in compliance with government regulations, then that is a cost that is simply borne.

If the cost to acquire land in a transaction is 1% of the value of the land in legal services, then due to extensive government compliance, that is simply the cost. If parties must therefore pay one thousand dollars in legal fees to pay for a five thousand dollar house, then that is simply the cost. That five thousand dollars might be made up of various title searches, submissions to government, review of the contracts, and so on. There might be a discretionary amount based on value that might go into negotiation or review of contract, as described above, but where there is simply some inefficiencies in the system, that is something that can be automated.

So if parties negotiate contracts on an individual basis as they do in NSW, which requires a solicitor to review the contract and negotiations to go backwards and forwards between solicitors for the vendor and the purchaser, then those costs might be overall higher than say SA, which typically uses a templated contract for most residential land purchases. Of course, the comments about the need for legal services still apply.

Accordingly, when there is a purchase of a ten million dollar shopping centre, the parties, both vendor and purchaser, would spend significantly larger amounts getting contract review and negotiation. However, at a minimum cost we can get efficiencies in the system. This is more about the adoption of automation, improved services, efficiencies and legal technology by government and regulating authorities, either direct regulating or institutions, rather than competition amongst legal services.