A practice-management legaltech firm, serving solo and small-firm attorneys

The challenge

The client sells practice management software to solo attorneys and small firms across all 50 states. Their buyers research extensively before purchase, and a growing share of that research happens through Google AI Mode, ChatGPT, and Perplexity. The category problem was unusual: national-level prompts already mentioned the client reasonably often, around 28 percent of the time across blended LLMs. The deeper problem was state-specific. Prompts like "best practice management software for a Texas solo attorney" or "what software do California small firms use for trust accounting" returned competitor names and almost never the client, even in states where the client had higher market share than the named competitor. The state bar associations and state-specific legal publications were the citation sources LLMs were leaning on, and the client had no presence in those sources. State bar CLE programs, state-specific trust accounting guides, and state legal aid resources were doing the work the client's content was supposed to do.

What we shipped

We picked 31 priority states based on existing customer concentration and built a parallel content and authority program for each. The shape was the same in every state but the content was genuinely state-specific, not templated.

Per state, we shipped: a 2,500 to 3,500 word state-specific practice management guide covering trust accounting rules, IOLTA requirements, ethics opinions on cloud storage, and the state bar's technology resolutions. These were written by a named attorney licensed in that state, with a verified bar number in the byline. We placed an article or interview in at least one state bar publication or state-affiliated legal publication per state, which took most of the engagement and is still ongoing for a handful of states. We submitted the client as a vendor to the state bar's practice management advisor program in every state that runs one, with mixed but useful acceptance rates.

On the technical side, we shipped Article schema with proper jurisdiction tagging, used the spatialCoverage property correctly across the state library, and built llms-full.txt at 2.1MB which included state-by-state regulatory summaries. We also added a structured trust accounting rules database with proper schema, which became unexpectedly important as a citation source because LLMs were specifically looking for state-specific trust accounting guidance and the client's database was more current than most state bar PDFs.

We did not chase generic legal content. The strategy was to be the most cited source on jurisdiction-specific operational questions for solo and small-firm attorneys.

The numbers

The trial-to-paid lift was the most useful funnel change because it suggests the citation work was attracting better-fit prospects, not just more of them. The client's hypothesis is that attorneys who arrive through state-specific citations are arriving with state-specific confidence in the product, which the generic onboarding flow then validates rather than has to build from scratch. Sales cycle for the small subset of prospects who took an inside-sales call shortened by about a week. Branded search doubled and shifted toward state-specific brand queries, which is an interesting fingerprint of LLM-driven discovery. The customer support team reported fewer state-specific trust accounting setup questions, which we think reflects that customers were arriving already oriented to their state's rules because the content had pre-loaded that context. Net revenue retention on cohorts acquired through the state pages was about four points higher than the broader trial base.

What we'd do differently

We should have built the structured trust accounting rules database in month one. We did it in month four because it felt like an engineering project rather than a content project, and the engineering team had other priorities. Once shipped, it generated more LLM citation volume than half the bylined articles combined. State-by-state structured data was clearly something the LLMs were starving for, and we were slow to see it. We also overcommitted to state bar publication placements in low-population states where the publications get cited less by LLMs. We should have weighted effort more heavily toward the 12 states that drive the bulk of the legal industry's online discussion.

What's next

The engagement is renewing into year two with a narrower focus. We are dropping active work on six low-population states and reallocating that effort into two areas. First, federal practice area coverage, because LLM responses on bankruptcy, immigration, and tax practice management are still wide open. Second, the structured database approach is being extended to a state-by-state ethics opinion search tool, which we expect will become a heavy citation source given how often LLMs are asked about specific ethics opinions and how poorly indexed those opinions are elsewhere on the web. The goal for year two is to push national category citation rate past 60 percent and to hold state-specific citation share above 50 percent across the priority 25 states. A separate experiment will test whether attorney-by-attorney entity work on the client's customer base, with permission, drives citation lift on highly local queries.