What changes when capacity stops being the constraint

A claims practice that automates assessment and preparation typically sees throughput per fee earner rise several-fold within a quarter. The interesting part is what happens next, because the constraint does not disappear — it relocates, and the firm is usually unprepared for where it lands.

When a file took two hours to assess, intake volume was never the problem. Once assessment takes minutes, the practice is bounded by how many qualifying claimants it can reach and onboard. Firms that automate preparation without rethinking acquisition end up with a highly efficient team waiting for work.

This changes who a firm needs to hire. The marginal useful hire after automation is frequently not another fee earner but someone who can build and run an intake channel — and partners who have spent a career managing legal capacity find that an uncomfortable reallocation.

Filing more cases means carrying more disbursements and more months of unrecovered work in progress. A practice that triples filings has tripled the size of the gap between outlay and award, and that gap is the most common reason growth stalls after a successful automation project.

Some firms close it with funding against filed cases, some by staging expansion to match recoveries, some by shifting fee structures. The mistake is not choosing one; it is not modelling the gap before the volume arrives.

At eighty active matters, a supervising partner can hold the state of the caseload in their head. At six hundred, they cannot, and informal oversight quietly becomes no oversight. Firms need sampling protocols, error-rate tracking by document type, and a defined escalation path long before they feel the absence of them.

The regulatory position is unchanged by automation: the lawyer of record remains responsible for every document filed. What changes is that discharge of that responsibility must be designed rather than assumed.

The tasks that survive automation are the ones requiring judgement under uncertainty: whether a marginal case is worth running, how to respond to an unexpected court order, when to settle. These are also the hardest tasks to train juniors on, and the traditional training ground — bundling, chronologies, first-pass review — is precisely what has been automated away.

Firms that have thought about this are deliberately routing juniors into supervised review of machine output and early-stage strategy work. Firms that have not are discovering in year two that their pipeline of capable seniors has thinned.

The discipline worth adopting before an automation programme starts is to name the next three bottlenecks and decide, in advance, what the firm will do when each arrives. Automation is not an endpoint. It is a change in which problem the firm is solving.