The Clash built a two-and-a-half-minute classic out of a single agony: a narrator standing in a doorway, unable to decide whether to stay or leave, aware that either choice carries a cost. If he leaves, there will be trouble; if he stays, the trouble doubles. He begs the other person to just tell him what to do — and, of course, the answer never comes from anyone but himself.
Every business client facing a dispute stands in that same doorway. Fight, or settle? Hold the ground, or walk away? And, increasingly, they arrive having already asked a machine for the answer. This post is about how to actually make that decision, and about what the new crop of AI case-prediction tools can and cannot tell you.
If I Go There Will Be Trouble, and If I Stay It Will Be Double
The hardest thing to accept about the litigate-or-settle decision is that there is no cost-free option. Litigation carries the obvious costs — fees, time, distraction, the risk of an adverse judgment, the years a case can steal from a business. But settlement has costs too, and they are easy to undercount: money paid on a claim you might have beaten, a signal to the next counterparty that you can be moved, a precedent within your own industry that you fold under pressure. Mick Jones’s instinct is right. Neither door is free. The task is not to find the painless choice; it is to price both doors honestly and pick the cheaper one over the life of the matter.
That pricing is not just about dollars. It runs through questions a spreadsheet does not capture: Is this a relationship you need to preserve, or a counterparty you will never see again? Is there a message you need to send to the market? Can you actually collect if you win? How much management attention will a two-year fight consume, and what is that attention worth deployed elsewhere? A number that ignores those factors is not an answer. It is a starting point.
So You Got to Let Me Know (pero que tienes que decir)
Here is what is new. Clients now walk in having run their dispute through AI-powered litigation-analytics tools — platforms that compare their facts against thousands of historical outcomes to produce settlement-value ranges, predict motions, and even score a particular judge's tendencies. The marketing around these tools is confident, with some vendors claiming eighty to ninety percent accuracy in forecasting outcomes, and at least one reporting that a firm using predictive valuation improved its settlement success and shortened its time to settlement.
Used well, these tools are genuinely useful. They impose discipline, they surface base rates a lawyer's gut might miss, and they can keep a client from over- or under-valuing a case out of emotion. The danger is treating a confident percentage as if it were the decision itself. These systems are strongest exactly where the data is thickest — ordinary contract and employment disputes with many comparable cases — and weakest precisely where the money usually is: the novel question, the thin-precedent fight, the matter whose outcome turns on how a particular witness holds up on cross. And no model captures the human dynamics of a case — the counterparty who will spend irrationally to avoid admitting fault, the judge having a bad month, the deal that quietly settles the moment real discovery begins, or the lawyer who is extraordinarily effective at advocacy.
This Indecision’s Bugging Me
Mick’s real problem is not that he chooses wrong; it is that he cannot choose at all, and the paralysis has its own price. Litigation punishes indecision. A client who will not commit to either a fight or a resolution tends to get the worst of both — running up fees in a case they are not actually prepared to try, then settling later from a weaker position and a lighter wallet. The value of good counsel at this stage is not a prediction dressed up as certainty. It is the judgment to look at the numbers, the AI estimate, the non-monetary stakes, and the realistic collectability, and then say plainly: this one is worth trying, or this one you should resolve now, and here is why.
That judgment is the part the tools cannot outsource. A model can tell you that cases like yours settle in a certain range. It cannot tell you whether yours is like those cases, whether your goals are financial or something else entirely, or whether you have the appetite and the resources to see a fight through. Those are conversations, not queries.
It’s Always Tease, Tease, Tease
The song never resolves; the narrator is still standing in the doorway when the music stops. Real disputes do not give you that luxury. The decision gets made, one way or another, and often the decision to delay is itself a decision. The clients who come through these moments best are the ones who price both doors honestly, use the AI estimates as one input rather than the verdict, weigh the costs a spreadsheet cannot see, and then commit — fully — to the path they have chosen. Hold your ground when the case and the numbers support it. Walk away when they do not. But decide with your eyes open, and do not let a machine, or your own indecision, decide for you.
This post is part of an ongoing series on how AI is reshaping the practical realities of litigation and the businesses that live through it. An earlier post took up the related question of when a dispute is worth litigating in the first place.
A Word About Silver Cain
Silver Cain PLC represents businesses in complex commercial and real estate litigation in Arizona and beyond. When Rebecca Cain and I founded the firm, we built it around direct partner involvement and senior trial-level judgment — the kind of counsel that will tell you plainly when a fight is worth having and when it is not. If the questions in this post are relevant to your business, or to the firms you retain, we are glad to have that conversation.
Leon Silver is an AV-rated trial lawyer at Silver Cain PLC, focused on commercial and real property disputes since 1989. Reach him at lsilver@silvercain.com.
