Segment 1 — The Cold Open

Segment 2 — Why It Costs What It Costs Today

Most routine legal services continue to be delivered through small practices or solo attorneys who price by the hour. A client pays for every minute spent reviewing facts, selecting clauses, and formatting the document even when ninety percent of the language is drawn from previously used templates. State bar rules that define the practice of law limit who may perform those steps, so tasks that could be handled by trained paralegals or software remain inside the billable envelope. Fragmented demand also plays a role: each matter arrives as a one-off engagement rather than as part of a high-volume standardized flow.

The result is that a simple lease or uncontested filing carries an effective price that reflects the carrying cost of the lawyer’s office, malpractice insurance, continuing-education requirements, and the need to spread fixed overhead across a modest number of matters per month. Clients experience the price as normal because the alternative—handling the matter without counsel—has historically carried its own risks and because transparent fixed-fee competitors have been slow to scale under existing rules. The baseline cost therefore feels like an unavoidable feature of professional judgment rather than a variable that could be attacked through different division of labor or different production methods.

Consider how the hourly model interacts with intake: an attorney must still spend time confirming that the client’s facts fit the template, even when the template itself has been validated in hundreds of prior matters. That confirmation step repeats across every new engagement because there is no shared, machine-readable data layer that would let the same facts be ingested once and reused. The carrying costs compound because each small practice must maintain its own insurance, trust-account compliance, and continuing-education credits, none of which scale with volume. When volume stays low, the per-document share of those fixed costs stays high, locking the price in place. This is why the current arrangement feels inevitable inside the profession even though the underlying document logic has not changed for decades.

Segment 3 — The Magic Wand Number & The Idiot Index

If a magic wand could instantly arrange the correct words into a legally sound document without any human time, the raw-material cost would be the electricity and storage required to hold and transmit a few kilobytes of text. That figure sits well below one cent per document and is already treated as negligible by any high-volume online service. The finished price of several thousand dollars therefore produces an Idiot Index on the order of hundreds of thousands to one.

The bulk of the gap lives in three places. First, the hours spent by a licensed attorney whose time is valued at three to six hundred dollars per hour. Second, the regulatory friction that prevents lower-cost providers from performing the same steps at scale. Third, the absence of standardized, machine-readable inputs that would allow the same document to be generated once and reused thousands of times with only minor variable data. Each of these layers compounds: an hour of review time is multiplied by the number of matters that cannot share a common workflow, and the regulatory barrier keeps volume low enough that automation investments never reach the scale seen in other document-heavy industries. The arithmetic is straightforward—an estimate of two thousand dollars divided by a sub-cent floor yields an index measured in the hundreds of thousands—yet the number itself is less important than the visibility it gives to where the cost actually resides.

To see the same arithmetic another way, imagine a single template for an uncontested divorce filing that has already been reviewed by counsel and accepted by courts in one jurisdiction. Once that template exists, the only variable inputs are names, dates, asset lists, and a handful of factual checkboxes. Those inputs can be collected through a structured form whose marginal cost is the electricity to run the server for a few seconds. If the regulatory environment allowed that filled form to be treated as presumptively valid, the remaining human step would be a quick exception check rather than a full reconstruction of the document. The difference between that exception-check model and the current hourly model is exactly where the hundreds-of-thousands-fold multiplier appears. Objections that every case is unique lose force once the data show that the overwhelming majority of routine filings fall within a narrow band of fact patterns already covered by published forms.

Segment 4 — The First-Principles Opportunity

A redesign would begin by separating the work into three layers: the legal logic itself, the data collection that populates the logic, and the final review or certification. The first layer can be encoded in decision trees or rule engines once and maintained centrally, removing the need to rebuild it for every client. The second layer can shift to guided questionnaires or direct data feeds that replace most of the intake interview. The third layer can be reserved for licensed review only on matters that exceed defined risk thresholds, while lower-risk outputs receive automated checks plus optional paralegal sign-off.

For these moves to reach the floor, three conditions must hold: the encoded logic must be accepted by courts and regulators as presumptively valid when it follows published forms; volume must be high enough for the fixed cost of building and auditing the system to be amortized; and unauthorized-practice rules must be clarified so that software plus limited human oversight does not trigger sanctions. The genuinely hard parts are the boundary cases—where facts deviate from the template—and the political question of how much protection the current rules are meant to provide versus how much they incidentally preserve an existing pricing model. Fixed-fee services that already publish prices in the low hundreds of dollars demonstrate that movement is possible; scaling them further would require the regulatory and data-standard changes above rather than further incremental efficiency inside the hourly model.

A concrete next step would be to publish the decision tree for one narrow matter type, such as a basic revocable living trust, alongside the exact court forms it produces. Regulators could then evaluate whether the output matches the form that an attorney would have produced by hand. If the match rate is high and the exception rate is low, the system could be granted a safe-harbor status that lets non-attorney operators run it under defined audit requirements. That single precedent would begin to separate the high-variance judgment work from the repeatable logic work, lowering the effective Idiot Index without removing licensed oversight where it actually adds value.

Segment 5 — The Lesson

When the dominant cost of a document is the calendar time of a highly credentialed professional rather than any scarce physical input, the opportunity lies in re-dividing the cognitive work so that judgment is applied only where variance is high. When rules written to protect clients also limit the reuse of correct templates across thousands of nearly identical matters, the first test of progress is whether those rules can be satisfied by auditable systems instead of repeated human hours. Tomorrow we will examine another domain where the same separation of logic, data, and certification could change the cost curve. Who will publish the first set of machine-checked forms that regulators treat as equivalent to traditional attorney work?

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.