There is a comfortable assumption in legal marketing that if a firm ranks well on Google, it is in good shape. We wanted to test whether that still holds now that a growing share of people ask an assistant instead of running a search.
So we scanned 84 North Carolina law firm websites. We did not pick struggling firms. Every single one was pulled from the first two pages of Google organic results for high-intent searches across criminal defence and DWI, family law, and workers’ compensation. These are the firms already winning at traditional SEO.
53 of the 84 firms scored under 60. Only 4 scored 80 or above.
Ranking well and being readable are two different jobs
Search engines rank pages. Answer engines recommend businesses, and a recommendation is a much higher bar. To name your firm, an assistant has to be confident about what kind of law you practise, where you practise it, who your attorneys are, and whether other people rate you. Most firms publish all of that in prose a human can read and a machine cannot reliably parse.
| What Google needs | What an answer engine needs | |
|---|---|---|
| The unit it returns | A ranked list of links | Two or three businesses, named |
| How it reads a page | Text, links and relevance signals | Structured facts it can lift with confidence |
| Practice areas | Inferred from headings and copy | Declared explicitly as Service data |
| Your attorneys | Bio pages it can crawl | Machine-readable records of who they are |
| Your reviews | Counted from Google’s own data | Invisible unless your site publishes the markup |
| Your questions | Nice for featured snippets | The format it quotes back to users most often |
Where the firms actually fell down
The pattern was consistent enough to be a little startling. The single most common gap was universal across the sample.
Not one firm in the sample had told the AI crawlers anything at all. Every robots.txt we read was silent on GPTBot, PerplexityBot and the rest, which means every engine is falling back to its own defaults and guessing at permission. It is a five-minute fix and nobody has made it.
The good news is that this is early
The best firm in our sample scored 92 out of 100, so this is clearly achievable with the budgets and the platforms firms already have. The top ten firms averaged 78.2 and the bottom ten averaged 36.7. That gap is not the product of one firm spending vastly more than another. It is the product of somebody having done a fairly short list of unglamorous things.
If you are the first firm in your county to do them, you are not competing for the machine’s attention with anyone. That window closes, and it closes market by market.
The honest limit
This measures whether a firm is readable by AI. It does not measure whether the firm is any good, and it is not a ranking of lawyers. A high score means the machines can describe you accurately. That is a marketing problem, not a legal one, and we would rather be precise about which problem we solve.