If you’ve spent time studying or working in the field of logistics, you’ve probably come across a concept known as the “last-mile problem.” Essentially, it means that in any process of delivering an object or a service to its destination, the very end of the process — the last steps required to complete the journey — is frequently the most difficult, complicated, and costly part.
The concept first arose in telecommunications, and that’s the area where most of us have actually experienced the last-mile problem. A giant network provider can assemble a massive and complex system of wires, electronics, and software programs to carry internet communications across great distances.
But to complete the connection process to you, the individual customer, an actual worker with a ladder and toolbag has to show up at your house and figure out the best way to run the appropriate cables in through your basement window or wherever. And when that worker arrives, they might find the natural connection point is actually inaccessible, or there's a weird feature in your house that creates unexpected difficulty, or you’re not home when they ring the doorbell. Each destination has its own idiosyncratic obstacles.
The same issue occurs in shipping. FedEx can load up a cargo plane with 1,000 packages and fly it from City A to City B swiftly and efficiently. But then those 1,000 packages need to get to their destinations, requiring 100 FedEx trucks that all have to navigate congested local streets, and then maybe the elevator in one office building is broken, so the driver has to take the stairs while their truck idles outside, blocking traffic.
It’s basically a problem of aggregation and disaggregation. You gain enormous speed and efficiency by moving a whole bunch of common elements towards a general destination, but then you give back most of those gains when making the specific connection required by each individual endpoint. That’s the logistics bottleneck known as the last mile.
More recently, a twin challenge has been identified at the other end of the process: the “first-mile problem.” If you live in a small town or exurban area far from City A and don't have a car, how can you get your package to the FedEx plane? This is an especially acute problem in public transportation: City A’s world-class light-rapid-transit system doesn’t hold much value for you if you can’t easily get to a train station. You have to reach the network before the network can help you.
The start and the end of a journey, therefore, create two different but related challenges. The first mile is about access; the last mile is about resolution. It probably won’t surprise you that I see an application here to the legal sector. I think AI will go a long way to solving law’s first-mile problem, but that lawyers will be indispensable to solving the last-mile problem — so long as we make two important adjustments right now.
As we know, access to justice is about more than the ability to hire a lawyer. Many people don’t recognize that a problem they're facing has a legal dimension and could feature a legal solution; many others do recognize this, but lack the practical ability to obtain the information, documentation, and direction they need to get started.
AI is already overcoming these obstacles for many people: identifying the legal issues in a person’s situation, explaining the relevant laws and procedures, and even drafting the basic legal documents needed to move forward. As legal AI systems become more proficient, they'll take people well past the first mile and into the “middle miles” that constitute most of the legal journey — bringing them closer and closer to a solution, overcoming the intimidation and inconvenience of the legal system and mitigating much of the information deficit people have always faced there.
It used to be that lawyers were required for almost the entire legal journey, from the first mile (access) to the last (resolution). With AI, a person can now make their own way well past Mile One (or Kilometre One, in my part of the world). So on a legal journey that covers a metaphorical 100 kilometres, a person could reach 60 km, or even 80 or 90 km, before they need a lawyer’s assistance. That's a simply extraordinary development, a turning point in the access-to-justice movement.
But eventually, in most matters of consequence or complexity, the person will need a lawyer. They’ll go as far as the AI can take them, but in almost all cases, they’ll come to a point where reaching their particular legal destination requires individual shipping and handling. Their package has been unloaded from the FedEx plane, or the telecommunications system has reached the neighbourhood control box, but now the final connections need to be made. And without those connections, the entire journey is practically useless.
The lawyer is the last-mile solution. The lawyer refines AI materials with expertise, nuance, and local context. The lawyer adds judgment and risk calculation to a suggested course of action. The lawyer provides verification and accountability for a legal strategy. The lawyer, as the saying goes, really does stand up in court or really does sit at the negotiation table and advocates for the client with poise, persuasion, and powerful rhetoric. There’ll be occasions where these elements aren’t required, where the entire legal solution process can be lawyer-free. For everything else, we'll need lawyers to perform the most important functions and make the final connections.
But lawyers will need to make two significant adjustments in order to successfully fulfil this role. And they both revolve around accepting that the nature of the client's legal journey has changed.
Lawyers are accustomed to meeting clients at the very beginning of their legal journey. The client realizes they have a problem that allows for or demands a legal solution; but they lack the expertise and wherewithal to obtain a solution on their own. This is the point at which the client has traditionally entered the lawyer’s office: burdened with an amorphous need, possessing neither the clarity nor the capability to take the first steps towards a solution, needing help with almost everything.
That will less often be the case in future. The lawyer is going to meet the client after the client has already travelled some distance down the road towards their solution, after having identified the nature of their problem and isolated the relevant laws and procedures and begun drafting the materials that will lead towards resolution. The client might be 20% of the way towards their goal, or 45%, or even 70% or more when they first show up. The lawyer will be playing catch-up to some degree.
In many cases, of course, the AI will have under-diagnosed the problem or missed an important regulation or pointed the client towards the wrong document. The lawyer will have to undertake some remedial work at the client’s expense. But that situation is likely to become less common in future, especially if and as institutional public-facing legal AI resources come online that provide more reliable and consistent guidance for the first part of the legal solution journey.
But even in these cases, the lawyer and client will not be starting from a blank slate. The yawning gap in knowledge and capacity that once separated lawyers and clients is narrowing by the day — and so is the power imbalance that the gap created. So lawyers have to start recalibrating their expectations about the confidence and sophistication of the people who come to them for assistance, because these features are only going to grow from here.
But the second adjustment is even more important: Lawyers have to recalibrate their pricing systems, too. When people needed lawyers for the entire length of their legal journey, a lawyer could simply charge for every step of that journey, every metre of the distance, every hour spent. AI is going to shorten the length of the journey for which the client requires the lawyer's intervention: the client might get 60% or 80% or 90% towards their destination before a lawyer’s presence is required.
But that last 40% or 20% or 10% that requires the lawyer's presence is significantly more valuable than the distance that came before it. If AI can bring a client 80% of the way to their solution for almost nothing, then the market price of travelling that 80% collapses.
The remaining distance, however, is essential: the expert evaluation, judgment, application, advocacy, and advice without which everything that came before it is useless — like a FedEx package sitting on the tarmac, or internet access locked away in a box on a power line. Solutions require a final connection. If you don’t plug in the cord, the lights won't turn on.
Billing by the hour is like billing by the kilometre. If you only charge the client for the final 10% of their journey, you are vastly underpricing the value of your services, because that final 10% requires your scarce and indispensable expertise. It’s in that last mile of the journey where a disproportionate share of value will be located. That’s why lawyer pricing has to shift away from measuring the amount of lawyer input into the process, and towards measuring the significance of the lawyer's output to the resolution.
There’s a little piece of Latin we all learned in law school that every lawyer needs to incorporate into their business model: sine qua non. “That without which not.” Recognize that the client won’t need you for much or even most of their legal journey anymore. But for those parts of the journey where you are needed, your value is exceptional — because without your essential contributions when it matters most, the client won’t reach their destination. If you apply the sine qua non principle to your legal pricing, you'll be very well positioned to manage the profound transition towards which the legal business model is headed.
People won't need lawyers to walk every step of their legal journeys anymore. Restructure your business and reorient your perspective to become a lawyer who walks the essential steps — the one who enables the final connection.
Hopeful and balanced. And terrifying. How do we lawyers who have been doing things for decades by the billable hour evolve to apply the sine qua non principle to our own legal pricing? Thank you for sharing your wisdom and inspiration; it is very much appreciated.
This is very insightful. An access to justice business model could involve an AI powered deep state specific legal information website, including county court information, with free calculators and decision tools, plus automated legal document assembly for legal forms and document, with automated document review for quality control, plus access to a curated network of law firms delivering limited legal services, (unbundled) for a fixed fee.