Friday reading for the rest of the week.

Litigation funders price risk against a cost base — and AI is compressing that cost base in real time. What happens to funding models, ATE premiums and costs budgets when the hours a case was budgeted to take are no longer the hours it takes? Closing Series 1, this piece follows the AI dividend into the funded-litigation market, where the question of who captures the saving is at its sharpest.

Pan-NOx was not decided by claimant numbers. It was decided by case theory, evidence, and the capacity to sustain both over years. The defendants concentrated the fight on the statutory definition and let the construction argument carry the volume of allegations away; Mercedes converted a systemic vulnerability into a discrete, time-bound point by fixing it a decade before trial. For anyone who funds, prices or manages group litigation: strategy and resources still beat scale

The general counsel slid a single sheet across the table. "That's your last four matters — hours billed against what our model says the work should take now the disclosure tool does what it does. The gap's about thirty per cent." For years pricing reform was debated almost entirely among lawyers while the billable hour carried on undisturbed. That period is ending — not because the arguments improved, but because clients stopped asking and started mandating.

"We looked at fixed fees," the practice head told me, in the tone people use for a holiday they nearly booked. The working group's paper is still on the shared drive; the firm still writes off the overruns it daren't pass on. The arguments for alternative fees were never the problem — the pressure was. AI has made the gap between time spent and value delivered impossible to keep out of sight. The alternatives work. The open question is whether a firm has the discipline to run