A criminal defense attorney I know recently mentioned that ChatGPT has become his best source of new cases. People ask about their situation, the AI cites his work, his name appears in the answer, and they treat that mention the same way they’d treat a recommendation from a trusted friend. They Google him, find his site, and book a consultation.
This isn’t rare anymore. Generative systems—whether inside search like Google’s AI Overviews or inside assistants like ChatGPT, Claude, Gemini and Copilot—are reshaping how clients discover attorneys. But they don’t read like humans. They slice content into pieces, evaluate each fragment, and decide which ones cleanly answer a specific part of a question.
If your writing matches that structure, your odds of being surfaced and cited rise dramatically. Publishing this piece on The Legal Examiner is itself an example of the strategy in action.
Use Query Fan-Out to Answer What AI Really Looks For
Ask a broad question, and an AI system immediately breaks it into sub-questions.
“What should I do after a rideshare accident?” turns into:
• which insurance policy applies
• what a passenger should document
• how liability works
• which deadlines matter
If your article only scratches the surface, you give the system nothing precise to extract. But when you address those sub-questions directly, you become a clean, ready source for citation.
Attorneys already know the real questions clients ask. Writing with those in mind aligns your content with the way these systems think.
Make Your Sections Passage-Ready
LLMs score passages, not pages. Every section stands alone. That means your headings need to be actual questions, not vague topics.
Weak: “Insurance Issues”
Strong: “Which Insurance Policy Covers a Rideshare Accident?”
And the body needs to shift from paragraph soup to structured, answer-ready content.
Before (paragraph soup):
“Rideshare accidents involve complex insurance questions because there are multiple policies that might apply depending on the driver’s status when the collision occurred, and Uber and Lyft have different coverage phases that activate based on whether the driver had the app on, whether they accepted a ride, and whether a passenger was in the vehicle, plus personal auto insurers often include rideshare exclusions that limit or eliminate coverage entirely.”
After (passage-ready):
Which Insurance Policy Covers a Rideshare Accident?
Coverage depends on the driver’s app status. Uber and Lyft use a three-phase insurance model.
Phase 1: App on, no ride accepted. Limited liability coverage applies.
Phase 2: Ride accepted until passenger exit. Full commercial coverage applies, typically up to $1 million.
Personal policies: Most carriers exclude rideshare activity whenever the app is active.
Clear structure. One idea at a time. Around 120 words. Easy for humans to read and easy for AI to extract.
Aim for 100 to 300 words per section. One question per heading. One idea per paragraph.
Where to Apply This Structure
This approach works everywhere: your firm’s site, bar association publications, guest posts on niche blogs, and broader industry outlets.
Some platforms, though, naturally amplify your authority. High-trust networks with strong editorial signals and established domain reputation tend to appear more often in generative outputs. Platforms like The Legal Examiner function that way. Publishing here as a contributor strengthens the clarity of your entity footprint and improves the likelihood your work will be cited by AI systems.
The attorneys who end up most visible across generative platforms publish consistently across multiple trusted surfaces. Each one reinforces who you are, what you practice, and why your explanations can be relied on. That pattern is what helps AI systems surface you and cite you, whether the answer appears inside AI Overviews or inside standalone assistants like ChatGPT or Claude.
