Most law firm marketing in 2026 is optimizing on what AI engines display. The smarter work is optimizing what AI engines retrieve before they display anything. The single highest-leverage piece of that retrieval-layer work is entity consolidation — making the firm and each attorney resolve as coherent, well-described entities across the source corpora the engines pull from. It is the work that compounds. It is also the work most agencies don't do, because it is invisible to traditional SEO dashboards.
This piece is for partners, marketing leads, and operators who have read the headlines about AI search and want to know what to actually do next. The answer is not "publish more content." The answer is "make sure your firm and your attorneys are the same firm and same attorneys everywhere AI engines look."
What "entity consolidation" actually means
When ChatGPT retrieves content for "best immigration lawyer Orange County," the model isn't searching the open web. It's pulling from a curated source corpus including Avvo, Justia, JD Supra, Google Business Profile, state bar member directories, court filings, and a small handful of other authority surfaces. Each of those surfaces has its own representation of "your firm" and "your attorneys." When those representations agree — same firm name, same NAP data, same attorney roster, same practice areas, same languages spoken, same admission years — the engine resolves them as a single entity. When they disagree, the engine treats the firm as ambiguous text and either picks one representation arbitrarily or excludes the firm entirely from the answer.
The technical term is entity resolution. It is the same process Google's Knowledge Graph uses, the same process Wikipedia disambiguation pages exist to support, and the same process retrieval-augmented generation systems run during the retrieval step. The practical effect for law firms is that consolidated firms get retrieved; fragmented firms don't. Every firm we audit that has high domain authority and low AI citation rate has an entity-resolution problem at the root.
A concrete example from a recent audit: a California firm whose business card lists `one domain` but whose actual website operates at `a different, unrelated domain` under the brand "a different brand." AI engines cite the firm consistently for local immigration queries — but as "a different brand," never as "the name on their card." A buyer who finds the card and searches for the card-brand discovers nothing; a buyer who asks ChatGPT discovers the website brand. The two channels are split. The fragmentation costs the firm citations on both names because the engines can't reconcile them as one entity.
The seven surfaces that need to agree
For California law firms specifically, seven surfaces define entity consolidation. Each requires its own attention; consistency across all seven is the goal.
Surface one: the firm's own website. Canonical firm name, address with consistent formatting (suite number style, street abbreviation style), phone number with consistent formatting, attorney roster including admission years, practice areas described in language that maps to the practice area taxonomy AI engines retrieve against. If the firm has a trade name and a legal entity name, the relationship between them is declared explicitly — typically with a "[Trade Name] is a trade name of [Legal Entity Name]" disclosure in the footer.
Surface two: California State Bar member directory. Each attorney's State Bar record shows their bar number, admission date, primary practice address, and current status. The practice address must match the firm site's address verbatim. Most attorneys haven't updated their State Bar address in years; the State Bar's record is the canonical legal authority, and it disagrees with the firm site more often than not. Login required: attorney logs in at apps.calbar.ca.gov and updates their record. Roughly 30 minutes per attorney.
Surface three: Avvo. Each attorney has an Avvo profile, claimed or unclaimed. Avvo auto-creates profiles for every licensed attorney, populating them with public data — admission year, practice area inferences, sometimes outdated contact information. An unclaimed Avvo profile with stale data actively damages entity resolution because AI engines retrieve Avvo content and the engines' answer reflects the stale Avvo record rather than the firm's canonical record. Claim every attorney's Avvo profile, populate the firm field with the canonical firm name, list current address, declare languages spoken, and add the practice areas the attorney actually handles. Each profile takes roughly an hour to claim and populate the first time; sustained maintenance is much lower.
Surface four: Justia. Justia maintains a separate lawyer directory and its own attorney profile system. The data Justia pulls from court filings and bar records may conflict with the firm site's representation. Claim each attorney's Justia profile (justia.com/lawyers), populate canonical fields, and crucially: bidirectionally link the Justia profile to the firm site's attorney bio. The bidirectional link is itself an entity-resolution signal — AI engines treat "site A links to site B and site B links to site A" as a stronger entity-equivalence signal than either link alone.
Surface five: Google Business Profile. GBP is the canonical local-business representation Google uses, and AI Overviews + AI Mode pull from it for any local-intent query. GBP needs: correct primary category (more specific is better — "Personal Injury Attorney" beats "Lawyer"), correct secondary categories, complete attributes (languages spoken, accessibility, parking, payment methods, online appointments), service area declarations, complete photos. Multi-location firms need one GBP per location; the locations must reference each other through the firm's structural data, not as separate businesses.
Surface six: Martindale-Hubbell and FindLaw. Both are part of the Thomson Reuters legal-marketing network and feed into AI engine retrieval for legal queries. Less individually impactful than Avvo or Justia, but together they form a meaningful portion of the citation surface for corporate, M&A, and complex commercial litigation queries specifically. Claim profiles, populate canonical data, cross-link.
Surface seven: JD Supra. Each attorney who publishes legal commentary should have a JD Supra author profile with the canonical firm name, current contact, attorney photo, and a populated bio paragraph. JD Supra's structured author metadata is exactly the kind of clean entity declaration AI engines reward. Even attorneys who don't publish heavily benefit from having the profile claimed and populated; the existence of the structured profile improves entity resolution even before any bylines accumulate.
Those seven cover the surfaces AI engines retrieve from heavily. A complete entity-consolidation pass touches all seven for every attorney in the firm. The work compounds — once a firm is a coherent entity across all seven, every subsequent piece of marketing work (content production, citation source acquisition, schema work) gets more leverage because the foundation is solid.
The operational pass — 40 hours, step by step
How to actually do entity consolidation in 40 hours of focused work, executable by a paralegal or marketing lead under attorney supervision.
Hour 1-2: Inventory pass. Open a spreadsheet. For the firm and each attorney, document every surface where they currently appear. Use the firm's name and each attorney's name as Google queries; document what shows up in the first three pages. Note inconsistencies: address differences, name suffix variations ("LLP" vs "LLC" vs "P.C."), missing attorneys on some directories, wrong primary practice areas, outdated phone numbers. The inventory is the baseline; subsequent passes measure against it.
Hour 3: Canonical-data definition. Decide on the canonical firm name (legal entity vs trade name), address format ("Ste 240" vs "Suite 240"), phone number format ("(714) 541-2400" vs "714-541-2400"), and attorney roster. Document any historical alternate names — predecessor firm names, "formerly known as" trails, trade name relationships. This document is the source of truth for every subsequent surface update.
Hours 4-25: Surface-by-surface update pass. The bulk of the work. For each of the seven surfaces, update to match the canonical. Avvo and Justia profiles get claimed where unclaimed; State Bar records get updated where outdated; GBP gets verified, recategorized, and fully populated; Martindale-Hubbell and FindLaw get claimed and populated; JD Supra profiles get created or updated. Each attorney's full pass across all surfaces takes roughly 2-3 hours; a firm with seven or eight attorneys completes this pass in about 20-25 hours.
Hours 26-32: Cross-linking pass. Where surfaces allow it, link them to each other and to the firm's own bio pages. The firm site's attorney bio links out to that attorney's Avvo profile, Justia profile, JD Supra author profile, State Bar lookup, and Martindale-Hubbell record. Each external surface, where it supports an external link field, links back to the firm bio. Bidirectional linking is the entity-resolution amplifier; one-way links help but bidirectional links signal "this attorney is the same entity in both places" far more strongly.
Hours 33-37: Schema markup on the firm site. Implement LegalService schema for the firm, Person schema for each attorney, ProfessionalService schema where appropriate. This is the firm's own canonical declaration of its entity structure, expressed in the machine-readable format AI engines use during retrieval. Important: schema must match the visible page content. Schema that declares facts the visible page doesn't support triggers Google's penalty for spammy structured data — and is detected by AI engines as low-trust signal.
Hours 38-40: Multilingual / language indicators + verification. If attorneys speak languages other than English, declare those on every surface that supports a "languages spoken" field. Avvo, Justia, JD Supra, GBP, State Bar — each has a place for this. Most firms only declare it on their own site; the AI-engine retrieval for non-English queries depends on the language declaration being multi-source. Final two hours: re-run the inventory and confirm consistency across all seven surfaces.
A worked example
One California firm Legal Torch recently audited is a useful illustration. The firm serves a Cantonese- and Mandarin-speaking clientele in San Francisco. Their firm-name search returns a coherent entity — the firm site, the State Bar registration, the Google Business Profile all agree on the basics. But the entity consolidation breaks down at three specific points:
Attorney inconsistencies. Three of the firm's attorneys have Avvo profiles with stale practice-area tags — Avvo's auto-inference categorized them as "Family Law" specialists based on a single case filing, when their actual practice mixes business litigation and family law. Each profile needs to be claimed and re-categorized.
Language declaration gap. The firm site declares "fluent in Cantonese and Mandarin" prominently. Avvo profiles for the attorneys don't declare any non-English language. Justia profiles don't declare any non-English language. GBP doesn't declare any non-English language. The AI engines retrieving for "Cantonese-speaking landlord-tenant lawyer San Francisco" don't have multi-source evidence that the attorneys actually speak Cantonese; the firm's language differentiator is invisible to entity resolution.
Chinese-language page disconnected from the entity. The firm has a `/chinese` page on the site that ranks position 88 in the US for the Chinese term 事務所 ("law firm"). The page exists and is indexed, but it isn't structurally connected to the canonical English-language attorney bios. AI engines see two separate but unlinked entities: an English-language firm with attorneys, and a Chinese-language page with no attorneys attached. Adding hreflang declarations, Person schema with `knowsLanguage` declarations, and explicit cross-links between the Chinese page and the English attorney bios resolves this in roughly 4 hours.
The expected citation-rate lift from a complete entity-consolidation pass for this firm: within 90 days, AI citation rate on Cantonese-language and Chinese-language legal queries roughly doubles. The work is bounded, executable in-house, and produces measurable results.
Why this is invisible to traditional SEO
Traditional SEO tools — Semrush, Ahrefs, Moz — measure: domain authority, backlinks, ranking keywords, organic traffic. None of those metrics capture entity consolidation. A firm can have Authority Score 20 and 1.2K referring domains and still be invisible to AI retrieval because the engines can't resolve the firm as a coherent entity. The traditional-SEO worldview doesn't have a metric for "is this firm a coherent entity to AI engines." Which is why traditional-SEO agencies don't do this work; their dashboards don't have a place to put it.
Semrush's relatively new "AI Visibility" metric is the closest existing measurement, but it samples broad queries and doesn't diagnose the underlying entity-resolution state. A firm with low AI Visibility may have an entity-consolidation problem (the most common case), a content-depth problem, a compliance-language problem, or all three. The metric tells you something is wrong; it doesn't tell you which thing. Diagnosing the entity-resolution layer requires the inventory pass above.
The structural finding is that entity consolidation is the highest-correlation, lowest-glamour, most-overlooked piece of AI-search work in legal-vertical marketing in 2026. It takes a focused paralegal or marketing lead about 40 hours to complete the full operational pass. The citation-rate lift typically appears within 60-90 days. The work compounds — once the firm is a coherent entity to AI engines, every other piece of AI-search work (content production, citation source acquisition, schema work) gets more leverage because the foundation under it is stable.
What this isn't
This isn't the entire AI-search playbook — it's the foundation layer. The other pieces matter and are covered elsewhere in this series: schema markup specifics (referenced briefly here), citation source acquisition like JD Supra and bar journal bylines (covered in detail in the citation-sources piece), Cal Bar Rule 7.1 compliance review on every output (covered in the Rule 7.1 piece and in the 2023 guidance piece), and the mechanics of how the engines actually pick which firms to recommend (covered in the mechanics piece). Entity consolidation is the layer underneath all of those — the firm's entity must resolve coherently before any of the other work pays off at full strength.
Bottom line
Entity consolidation is foundational AI-search work that traditional SEO doesn't measure, most agencies don't do, and most firms haven't audited. The investment is bounded — roughly 40 hours of paralegal or marketing-lead time, plus a few hours of attorney attention for State Bar and bar-association updates. The return is durable: a firm that resolves as a coherent entity across the seven surfaces above will be cited more reliably across ChatGPT, Perplexity, Google AI Overviews, and Claude than a firm with twice the content production but fragmented entity representation.
If you want to see your firm's current entity-consolidation state — across Avvo, Justia, GBP, State Bar, JD Supra, and the other surfaces that matter — the free AI Visibility Audit covers the inventory pass. The full operational pass is delivered as part of the VERDICT retainer's Entity pillar, with progress tracked surface-by-surface in the monthly PROOF Report. Either path: start with the inventory. You can't fix what you can't see.
New to the topic? Start with the complete guide: GEO for Law Firms — what it is, what it costs, and what the citation data actually shows.
