Order of Operations

Every firm has decades of work product scattered across SharePoint, Outlook, a shared drive nobody has pruned since 2014, and three partners’ memories. Consolidating it into something retrievable is pillar two’s obvious half: the briefs, the templates, the winning arguments, the deposition outlines, the intake checklists, the CLE material, the standard answers to the standard questions.

The half firms skip is the outcome data. Not the documents, the facts about what happened. What this matter type actually costs to run. How long it actually takes, end to end, including the waiting. What it settled for, in front of which judge, against which carrier, with which fact pattern, at which stage. That is the asset with real economic value, and it is the one that exists nowhere until you decide to capture it, because no system you own was designed to record it.

Notice the loop back to pillar one. You cannot price a fixed fee responsibly without this. Everyone who has ever lost money on flat work lost it because they priced from memory instead of from a distribution. The outcome layer is what turns pillar one from a gamble into underwriting.

Governance goes in at the foundation

Not bolted on in year two. Before the first document lands: who can retrieve what, what your ethical walls look like inside a retrieval system that does not understand walls by default, what leaves the building and under whose terms, how long things are kept, and what gets verified before it reaches a client. ABA Formal Opinion 512 sets the floor on competence, confidentiality, supervision, and fees. Your own state bar’s opinions sit on top of it and are frequently narrower, so read yours rather than the national summary.

Search is becoming a conversation held without you in the room. The question is no longer whether you rank for “Jacksonville probate lawyer.” It is whether, when somebody asks an assistant what to do about their mother’s estate, your attorney is part of the answer.

Findable is table stakes: consistent entity data, structured markup, clean and identical profiles everywhere, an attorney who resolves to the same person across every source, questions answered in the language clients actually use rather than the language the statute uses.

Citable is the real work: original research, named experts, third-party corroboration. This is where the book goes, and the book is not the asset. A self-published title nobody references moves nothing. What moves a machine’s association between a problem and a person is the citation graph the book produces, the press that covers it, the podcasts that book the author, the bar journals that review it, the sites that link to it as a source. Write the book to get cited, not to have written a book. An annual report with real proprietary numbers in it often gets cited harder than a book and takes a tenth as long.

Callable is the frontier, and it is where pillar two pays off a second time: eligibility checkers, deadline calculators, jurisdiction-specific lookups, case-readiness assessments, document analyzers, intake endpoints. As assistants shift from answering questions to taking actions, you want to be something they can invoke, not only something they can quote.

Firms should not build their practice management system, their document management, their accounting, or their email. Those problems are solved, and the solved version is better than yours will be. Anything that is a commodity stays a commodity.

What is worth building is the narrow thing that is genuinely yours and genuinely absent from every vendor’s roadmap: the workflow you run two hundred times a year that no product models. A medical treatment tracker. A beneficiary document-collection flow. A case-readiness scorer. A referral routing portal. These are usually thin, a few screens over the data you started capturing in pillar two, and the thinness is the point.

Here is the part nobody writing these frameworks will tell you. Most internal software dies. Not some of it. Most. It gets built by an enthusiastic partner, used by four people, abandoned when the champion gets busy, and maintained by nobody. That is not an argument against building. It is an argument for writing the kill criteria down before the first line of code, and for sizing the budget around the death rate rather than around the success story.

That is what makes this an R&D function rather than a project. A fixed monthly number, a named owner, and a quarterly cycle: one expensive workflow, one bad client experience, one new distribution channel, one product idea. Prototype all four. Kill three on purpose.