Tracking when ChatGPT cites your law firm in 2026 means running three workflows in parallel: a weekly prompt library against ChatGPT, Claude, Perplexity, and Google AI Mode, a GA4 referral filter for AI traffic that actually clicks through, and a citation log that ties each mention back to the underlying source article or schema block. The DIY floor costs nothing but two hours of partner time a week. The paid floor starts at $29 per month with Otterly.AI, scales to $295 with AthenaHQ, and tops out at $499 with Profound. Most law firms either skip this entirely or buy a tool and never run the prompts, which is the same outcome dressed up in invoicing.
Why ChatGPT citation tracking matters more for law firms than for almost any other vertical
Three reasons. First, legal buyers are high-research, high-intent. A managing partner search for “best commercial litigation firm Dallas” used to start on Google; in 2026 it starts on ChatGPT 60 to 70 percent of the time among under-50 buyers, based on the public Pew and Higher Visibility data for 2025-2026. Second, legal services have huge lifetime value per client, so a single citation that pulls a buyer into a discovery call can return ten times the cost of a year of AI search tooling. Third, the bar for citation is high but specific: AI engines pull from Avvo profiles, Martindale-Hubbell, state bar pages, named-attorney bios with proper Attorney schema, and Tier 1 legal press (Above the Law, ABA Journal, Law360, Bloomberg Law). If you do not know which of those sources is feeding which AI engine, you cannot fix the gaps.
The mistake we see most often: a firm pays for an AEO tool, runs the dashboard once, and treats the screenshot as a deliverable. That is reporting, not tracking. Tracking is a closed loop where every cited prompt feeds back into content, schema, and outreach the same week.
What “citation” actually means inside ChatGPT, Claude, Perplexity, and Google AI Mode
The four engines surface citations differently, and the difference matters for how you track them.
ChatGPT (the consumer product and ChatGPT Search) cites web pages with footnoted source links beside the relevant claim. As of mid-2025, OpenAI began appending utm_source=chatgpt.com to desktop citation links, which means a click from the ChatGPT answer now passes a parameter you can filter on in GA4. Mobile and some legacy clients still strip the referrer, so a meaningful slice of ChatGPT traffic still lands as Direct in GA4. Treat the chatgpt.com referral row as a floor.
Perplexity has cited sources from day one and is the easiest engine to track because every claim has a numbered footnote. Perplexity Pro and the API both expose citations in structured form, which is why most AEO platforms (Profound, Peec AI, Otterly, AthenaHQ) treat Perplexity as the canonical reference engine.
Claude (Anthropic) cites sources when it uses web search through claude.ai, but its answers more often synthesize without inline links. For a law firm, the practical question is whether Claude names you in an answer about “top estate planning attorneys in Charleston,” not whether it links to your bio. Citation tracking for Claude is mostly mention tracking with light source attribution.
Google AI Mode and AI Overviews cite both URLs and direct entity callouts. Google’s citations skew heavily toward existing top-ranked organic results plus knowledge graph entities, which is why your standard local SEO and Google Business Profile work doubles as Google AI Mode optimization.
Tracking those four engines uniformly means logging the prompt, the engine, whether you were mentioned, whether you were cited (with a clickable source), what source the engine pulled from, and what competitor was mentioned in your absence. Five columns. Every week.
The DIY workflow that costs nothing but two hours a week
If you are a solo partner or a small firm not ready to spend $300 a month on tooling, this is the floor. The same workflow scales up once you add a paid tool.
Build the prompt library
Start with 40 prompts split across five categories:
- Practice area plus city. “Best estate planning lawyers in Charleston SC.” “Top construction defect attorney Dallas.” Ten prompts. These are top-of-funnel discovery queries.
- Practice area plus pain point. “How do I find a lawyer for a wrongful termination case in Texas.” “Who handles trust litigation in South Carolina.” Ten prompts. These mirror real buyer language.
- Named attorney. “Tell me about [your managing partner’s name].” “What cases has [your partner] argued.” Ten prompts. These test whether the AI engine has built a coherent entity profile of your team.
- Comparison. “[Your firm] vs [competitor firm].” “Which is better, [your firm] or [competitor firm], for personal injury.” Ten prompts. Comparison prompts are how late-stage buyers actually decide.
- Question-style AEO prompts. “What should I look for when hiring an antitrust attorney.” “How much does a corporate litigator cost in Atlanta.” Five to ten prompts. These are the FAQ-style queries where strong FAQ page schema wins.
Store the list in a Google Sheet with columns for: prompt, engine, date run, mentioned (Y/N), cited with link (Y/N), source URL, competitor mentioned, notes.
Run the prompts weekly
Pick a recurring 60-minute block. Open four browser tabs: ChatGPT, Claude, Perplexity, Google AI Mode (or the AI Overviews layer inside Google Search). Paste each prompt into each engine in a fresh chat (no memory carryover for ChatGPT and Claude; toggle off custom instructions). Log the result.
Two operational notes. First, run prompts in incognito or with a clean session, because logged-in profiles bias answers heavily. Second, geo-locate when relevant. ChatGPT’s answers for “best DUI lawyer near me” shift based on the IP it sees. Use a VPN if the firm has multiple offices.
Analyze, do not just collect
Three filters every week:
- Citation gaps. Prompts where you used to be cited but no longer are. These are the most urgent. Usually they signal a content refresh need or a competitor publishing fresher material.
- Mention without citation. Prompts where the AI names you in passing but does not link or quote you. These usually need a stronger schema layer on the relevant page or a press placement to give the engine a fresh source.
- Competitor wins. Prompts where a specific competitor consistently shows up in your absence. Reverse-engineer the source the engine cited. If it is a press hit, that is your next pitch target. If it is an Avvo profile, that is your next review platform play.
The whole workflow takes a partner or marketing lead two hours a week. That is the floor. Below that, tracking stops being a system.
The paid tools that actually justify their price for law firms
The AEO citation tracking market has consolidated to roughly six tools that matter. For law firms, three are worth shortlisting.
Otterly.AI starts at $29 per month for the Lite plan, which fits a solo or two-partner firm. You get weekly automated runs against ChatGPT, Perplexity, and Google AI Overviews. You lose Claude coverage, deep source attribution, and competitor benchmarking. The value is the automation: you stop forgetting to run the prompts.
AthenaHQ sits at $295 per month self-serve ($95 for the first month), with an enterprise tier on request. Athena’s Citation Engine (ACE) blends prompt auditing with proprietary prompt-volume estimates, which means it ranks your prompt library by how often real buyers actually use those queries. For a mid-sized firm with five to fifteen attorneys, this is the sweet spot. AthenaHQ also exposes source-level attribution, so you can see exactly which press article or directory profile triggered each citation.
Profound starts at $499 per month, scales to enterprise pricing, and raised a $96 million Series C in February 2026 at a $1 billion valuation. For a national firm or a firm with multi-state offices and a marketing team that can act on data, Profound is the category leader. It tracks all four engines, surfaces share-of-voice against named competitors, and integrates with Salesforce and HubSpot for funnel attribution.
Peec AI at €89 per month is the European mid-market equivalent of Profound. We use Peec AI on a couple of clients who wanted GDPR-clean data routing. The tool is solid; the U.S. integration ecosystem is less mature than AthenaHQ’s.
Two tools we no longer recommend for legal: any “AI rank tracker” that only tracks ChatGPT mentions without source attribution, and any generic media monitoring tool (Meltwater, Cision, Muck Rack) that bolted on an AI module in late 2025. Meltwater’s GenAI Lens is real and is improving, but the legal-specific citation depth is not yet there.
How to set up GA4 to catch the AI traffic that actually clicks through
Citation tracking tells you who is being mentioned. GA4 tells you who is converting. Two filters bridge the gap.
Custom channel group for AI sources
In GA4, go to Admin, then Channel groups, then create a custom channel group called “AI Search.” Add these source patterns: chatgpt.com, chat.openai.com, perplexity.ai, gemini.google.com, claude.ai, copilot.microsoft.com, phind.com, you.com. Save.
You now have a clean traffic acquisition row that shows AI referral sessions, conversions, and revenue side by side with Organic Search and Direct. For a law firm, the conversion event is usually a contact form submission, a phone call (if you have CallRail or similar piping events into GA4), or a ROI calculator submission.
UTM tagging on your own AEO content
Whenever you publish an FAQ block, a press release, or a deep-dive blog post you expect AI engines to cite, append a UTM-friendly canonical URL strategy. We use ?utm_source=aeo&utm_medium=organic&utm_content={slug} on every internal link inside AI-targeted content so we can see which exact pages are pulling AI clicks through.
Watch the Direct line
Direct traffic for law firms in 2026 contains a lot of stripped-referrer AI clicks, especially from mobile ChatGPT and Claude. The diagnostic: if your Direct traffic spikes the same week you get a Tier 1 press hit or a verifiable AI citation, the spike is almost certainly AI referral traffic that lost its referrer. Cross-reference with your prompt library log.
What to do with citation data once you have it
Tracking is upstream of action. Five action loops we run for law firm clients.
Cited and converting. Lean in. More content on the same topic, more variations of the same FAQ, more press pitches that link to the cited page. The engines are telling you what they trust about your firm. Feed it.
Cited but not converting. The AI engine likes you; the visitor does not. Audit the landing page. Bounce rate above 75 percent, no clear CTA, or a slow mobile experience kills the conversion. Half the law firm landing pages we audit do not even have a phone number above the fold.
Mentioned without citation. The engine knows you exist but does not have a link to point at. Build the link. Most often this means a deeper FAQ page on the topic, a longer named-attorney bio with full credentials and Attorney schema, or a press placement that becomes the engine’s preferred citation.
Competitor cited in your absence. Source-trace the competitor citation. If the source is Above the Law, your next pitch is Above the Law. If the source is the competitor’s blog post, your version of that post should ship in 30 days and be stronger.
Citation lost. This is the urgent one. A citation you had in March that you do not have in May means the engine found a fresher or higher-authority source. The fix is usually a content refresh with a new publish date and new substance, not just a date stamp. AI engines weight freshness, but they also detect lazy republishing.
How often should a law firm run this workflow?
For most firms, weekly. For a firm spending $5K+ a month on marketing or paying any agency for AEO, weekly is the floor. For a small firm running this in-house, biweekly is acceptable. Monthly is too slow; the AI engines update their training cuts and retrieval indexes faster than monthly cadence catches.
If you are running an internal AEO project, build the workflow into your operations calendar the same way you build in CLE deadlines. It is recurring, non-negotiable work.
FAQ
How much does ChatGPT citation tracking cost for a law firm?
The DIY floor is zero, with two hours of partner or marketing lead time per week. Paid tools start at $29 per month (Otterly.AI), step up to $295 (AthenaHQ self-serve), and reach $499+ (Profound). Most mid-sized firms land at the AthenaHQ tier.
Can I track ChatGPT citations in Google Analytics?
Partially. GA4 captures click-through traffic from ChatGPT when the referrer is preserved (desktop, since mid-2025, with utm_source=chatgpt.com). Mobile and stripped-referrer traffic lands in Direct. Use a custom channel group in GA4 plus a weekly prompt library log to triangulate.
Which AI engine matters most for law firms in 2026?
ChatGPT by volume, Google AI Mode by intent quality (because Google Search still dominates legal queries), and Perplexity for ease of tracking. Claude is growing fastest among professional services buyers but cites less consistently. Track all four; weight ChatGPT and Google AI Mode for action priority.
How long does it take to see ranking improvements after fixing AEO gaps?
For a published content fix or schema update, expect ChatGPT to surface changes in two to six weeks. Perplexity updates within days. Google AI Mode tracks Google’s organic index, so it follows your standard SEO timeline. For a press placement that creates a new citation source, expect AI engines to pick it up within four to eight weeks once Tier 1 publication has been indexed.
Should our firm hire someone to run this or do it in-house?
Most partners do not have the bandwidth to run a 40-prompt weekly library themselves. Either dedicate a marketing coordinator (40 to 60 minutes a week is plenty), have your fractional CMO add it to the scope, or hire an agency that builds it into your retainer. The tool by itself solves nothing; the workflow is the value.
If you want a working baseline of where your firm currently shows up in ChatGPT, Claude, Perplexity, and Google AI Mode, you can run a free version of this audit yourself with the prompt library above, or use our ROI calculator to see what fixing the citation gaps would be worth to your firm. For deeper help, book a call and we will run a full citation audit on your firm and your top three competitors before we ever talk pricing.
Tagged