The Definitive 2026 Playbook

By Robert (Bob) Levin, CTO of Mediate Lawsuit and Founder of Transformational Outsourcing

Last updated: 06/01/2026

How AI Engines Now Decide Which Neutrals Get Recommended — and the 90-Day Plan to Become One of Them

Last month, an in-house counsel at a Fortune 500 company opened ChatGPT and typed: “Who is the best employment mediator in Atlanta?”

The model responded in four seconds. It named three neutrals. It cited five source pages. It gave a confident, plausible-sounding recommendation.

Two of the three Atlanta mediators it named are well-credentialed, court-rostered, and panel-listed with AAA. The third has been retired for over a year.

Hundreds of Atlanta-area mediators were not mentioned at all.

The truth most ADR practitioners have not yet faced is this: the process by which mediators and arbitrators get discovered and recommended has changed, and it has changed quickly.

Gartner has projected that traditional search volume will fall by roughly a quarter over the next year as users shift toward AI-powered answer engines. Google’s AI Overviews now reach more than two billion monthly users. ChatGPT serves roughly 800 million users weekly. Perplexity, Claude, Gemini, and Copilot each process tens of millions of additional queries every week.

When an attorney needs to find a neutral, fewer of them are scrolling through Google’s ten blue links than was true even eighteen months ago. More of them are asking an AI engine a direct question and trusting the answer.

In traditional search, the competition for visibility ran among the top ten organic listings. In AI search, the competition runs among two to seven sources the model chooses to cite in a single response. The competition is tighter. The stakes are higher. And the rules are different.

For the ADR profession specifically, the rules are different in ways the broader generative engine optimization (GEO) literature has not yet engaged with seriously. Mediators and arbitrators are not law firms. 

The selection logic that applies to a personal injury practice or a real estate firm does not apply cleanly to a neutral. 

The signals AI engines weigh, the source hierarchies they trust, and the citation patterns they reward all behave differently in our profession than in adjacent ones.

This article is the first comprehensive treatment of that difference. It introduces a framework — the CITE framework — for how AI engines establish trust in a neutral. 

It provides a tested 90-day implementation plan for solo and boutique-firm practitioners. It addresses the platform-mediated authority problem honestly. And it announces the launch, in Q4 2026, of The State of Mediation — an annual benchmark of digital visibility patterns across the profession.

What follows is built for working neutrals. Most of the tactics can be executed without a marketing agency. The frameworks are yours to use. The goal is simple: that when an attorney or a party asks an AI engine for someone like you, your name is one of the two to seven that come back.

What GEO Actually Means for Mediators and Arbitrators

Generative engine optimization is the practice of structuring your digital presence — your bio, your credentials, your published work, your third-party citations, your schema — so that AI-powered search platforms can retrieve, verify, and recommend you when they answer a question about neutrals.

The phrasing matters. AI engines do not rank you. They retrieve you. They reason about you. They decide whether to name you.

The distinction from traditional SEO is important and frequently underappreciated. Traditional SEO optimized for position — the goal was to appear higher in a list of ten links. GEO optimizes for citation — the goal is to be among the handful of sources the AI engine chooses to cite when it gives a direct answer.

This shift changes the economics of visibility for neutrals in three ways.

First, the field of competition shrinks. A typical AI response to a question about mediators or arbitrators' names between two and seven neutrals or firms. 

That is a smaller, more concentrated competition than the ten organic search results of the previous era. The neutrals who win the citation game capture a disproportionate share of attention. The neutrals who lose it become structurally invisible.

Second, the trust signals shift. Traditional SEO rewarded a wide range of activities: link building, keyword optimization, content volume, and page speed. GEO rewards a narrower and more demanding set of signals: factual specificity, third-party verification, entity coherence across sources, and demonstrable evidence of expertise. 

The tactics that worked for the early SEO era — keyword density, link networks, content farms — actively hurt you in the GEO era. AI engines are designed to detect and discount them.

Third, the implicit endorsement compounds differently. When Google ranked you third in organic results, attorneys understood that ranking as the outcome of an algorithmic process. 

When ChatGPT names you as the recommended employment mediator in your city, attorneys read that as a recommendation — even though it is also algorithmic. The difference is psychological but consequential. Being cited carries a weight that being listed never did.

For mediators and arbitrators, the question is no longer whether AI engines will mediate discovery. They already are. The question is whether you have engineered your digital footprint so that when the question comes, the answer includes you.

This is what GEO actually means for our profession. Not page speed. Not keyword density. Not blog post volume. The work is to make yourself citable — by AI engines, on behalf of the attorneys and parties asking about you.

The next section explains why that work is structurally harder for neutrals than for almost any other professional services category — and why most generic GEO advice will fail if you apply it without adjustment.

The Four Anomalies That Make ADR-GEO Different

Most professional services — law firms, consultancies, medical practices — operate under a relatively standard set of digital visibility conditions. The firm is the entity buyers retrieve. 

The credentials are nationally standardized. The work product is publishable. The selection cycle is frequent enough that brand recall accumulates.

ADR breaks all four of these assumptions. Mediators and arbitrators operate in a structurally unusual digital environment, and any GEO strategy that ignores the unusual features will produce mediocre results no matter how competently it is executed.

The four anomalies, in order of how much they shape the practical work:

Anomaly 1 — The Dual-Entity Problem

In almost every other professional services category, the firm is the unit of selection. Attorneys hire “Jones Day.” Patients see “Mayo Clinic.” Consumers buy “McKinsey.” AI engines, built to favor coherent organizational entities, retrieve these names cleanly.

ADR works differently. Attorneys and parties do not hire JAMS or AAA in the way they hire a law firm. They hire a specific neutral — Judge X, Mediator Y, Arbitrator Z — and the institution is the channel through which the neutral is engaged, not the unit of selection itself. 

The human is the entity. The firm or panel is the context.

This creates a structural problem for AI retrieval. The systems are optimized to recognize, verify, and cite organizational entities. They have to work harder to recognize, verify, and cite individual humans. 

The neutral whose firm has a strong digital presence but whose own bio is thin is being recommended as their firm, not as themselves. That is not the same thing as being cited.

The work, then, is to engineer dual-entity authority: a firm or practice entity that is coherent and well-cited, paired with an individual-neutral entity that is independently retrievable. Both have to be built. Either alone is insufficient.

Anomaly 2 — The Platform-Mediated Authority Problem

The provider rosters that confer the most institutional authority on a neutral — AAA, JAMS, FedArb, NAM, CPR — are, simultaneously, surfaces over which the neutral has limited control.

This is not a criticism of the providers. The curation discipline that makes a roster valuable is exactly what limits how much an individual neutral can customize their bio page on it. Roster pages are constrained by design — for fairness, brand consistency, and operational scale. 

Most provider-platform bios offer limited outbound link permission, limited schema customization, and limited content depth. The information presented is enough to confirm credentials and direct an inquiry. It is not enough to function as a fully optimized AI-retrieval surface.

The consequence is that the strongest authority source for many neutrals — their panel membership — is paired with a weaker authority surface for their individual digital identity. 

A neutral with twenty years of AAA panel membership and a deep specialty record may still be undercited by AI engines because the AAA page that confirms their authority is structurally light on the kind of detail that drives retrieval.

The answer is not to leave the panels. It is to build owned-authority surfaces in addition to roster status. Your panel page handles institutional verification. Your own site, properly built, handles depth, specificity, and citation density. Both are needed. Neither replaces the other.

Anomaly 3 — Low-Frequency, High-Trust Purchase Under Confidentiality Constraint

An attorney selects a mediator perhaps a handful of times a year — sometimes once. The selection is consequential. The selection requires trust that exceeds what a routine vendor decision demands. 

And, uniquely in our profession, the selection is constrained by confidentiality rules that prevent the case-study content most professional services rely on to demonstrate expertise.

Mediators cannot publish detailed write-ups of cases they handled. Arbitrators in commercial proceedings often cannot publish the awards they issued. The content categories that drive trust signals in other professional services — “Here is the case we won, here is what we learned, here is the methodology” — are largely unavailable to neutrals.

This constraint is real, and most neutrals stop here. They conclude that the rules of their profession prevent them from producing the kind of content that drives visibility, and they accept lower visibility as a cost of practice.

That conclusion is wrong, and accepting it is one of the most expensive mistakes in ADR practice today. Confidentiality does not prevent content. 

It forces a better kind of content — pattern-level analysis, procedural commentary, framework explanation, hypotheticals, and methodology. Done well, this kind of content is more retrievable than case studies because it is more abstract, more reusable, and more obviously expert.

We will return to this point in the section on the Published-Reasoning Doctrine, where it becomes one of the most consequential strategic levers in the playbook.

Anomaly 4 — Credential Chaos

The legal profession has the bar. Medicine has board certification. Finance has FINRA, the CFA, and the CPA. Each is a single national or jurisdictional authority with clean, machine-readable credentialing.

ADR has nothing comparable. “Florida Supreme Court Certified Circuit Civil Mediator” is a precise credential. “Experienced commercial mediator” is meaningless marketing language. “Member of the AAA Commercial Panel” is verifiable. “Senior mediator” is not. 

The space between credentialed and uncredentialed runs through a wilderness of self-declared titles, retired-judge designations, certificate-program completions, and provider-panel memberships with widely varying standards.

AI engines, faced with this chaos, do exactly what they are trained to do: they default to recognizable, structured, third-party-verifiable signals when they can find them, and they ignore or discount signals they cannot verify. 

A neutral whose credentials are presented in unstructured prose, without verification anchors, will lose retrieval competition to a neutral with the same actual qualifications presented in structured, verifiable form.

This is the most exploitable of the four anomalies. The neutrals who present their credentials in machine-readable, third-party-anchored form are dramatically over-cited relative to their objective qualifications. 

The neutrals who present the same credentials in paragraph prose are dramatically under-cited.

The CITE framework that follows is built around responding to all four of these anomalies in a structured way. Each anomaly maps to one or more of the four CITE conditions. Each condition has a tactical execution path. The 90-day plan in the second half of this article walks the path step by step.

The CITE Framework: Four Conditions for AI Citation

The four anomalies in the previous section explain why GEO is structurally different for neutrals. The CITE framework explains what to do about it.

CITE stands for the four conditions an AI engine evaluates before naming a neutral in a recommendation. Each is a standard, not a phase — meaning the neutral who satisfies all four can be cited by engines today, and will remain citable as the underlying technology evolves. 

The standards survive tactical change. The specific tactics that satisfy them will not.

C — Coherence. Does the same neutral identity — name, credentials, jurisdiction, specialty — appear consistently across every authoritative source where you are mentioned?

I — Identification. Are your credentials, specialties, and case-type signals presented in specific, structured, machine-readable form rather than in vague prose?

T — Triangulation. Do at least three independent, authoritative third-party sources confirm the central claims of your practice?

E — Evidence. Is there published material — articles, methodology, reasoning, awards, framework commentary — that demonstrates how you actually think and work, not merely that you exist?

A neutral who meets none of the four is functionally invisible to AI engines. A neutral who meets one or two will appear inconsistently — cited by some engines on some queries, missed by others. 

A neutral who meets three is competitive in their geographic and specialty lane. A neutral who meets all four will be cited disproportionately, including on queries where their competitors have stronger nominal credentials.

The disproportion matters. AI engines, when faced with a roster of similarly credentialed neutrals, do not rank by raw credential strength. They rank by retrievability. 

The neutral whose digital footprint is engineered to be retrievable wins repeatedly against the neutral whose footprint is stronger in absolute terms but weaker in structure. This is the most counterintuitive finding in the entire ADR-GEO landscape, and the most important.

Each of the four conditions deserves unpacking. The 90-day plan in the second half of this article operationalizes them. Here, we explain what each one actually means.

C — Coherence

Coherence is the condition AI engines evaluate first. Before any engine decides whether to cite you, it has to be confident that the “you” it has assembled from across the web is the same person. 

If your LinkedIn says you are “Robert J. Smith, Mediator and Arbitrator,” your AAA panel page says “R. James Smith, Esq.,” your firm bio says “Bob Smith,” and your court roster listing says “Robert Smith, FL Cert. Circuit Civil,” the engine is now reasoning about what may be one neutral or what may be four similarly-named people. 

The engine resolves this ambiguity by becoming cautious. Cautious means undercited.

Coherence is built across several axes simultaneously. The name as it appears in every authoritative source. The professional title and credentials as they are stated. The jurisdiction. 

The specialties claimed. The headshot used. The contact information. The firm or platform affiliations. When all of these are consistent, the engine builds a confident entity profile. When they conflict, the engine fragments.

The most common coherence failures, in order of how badly they hurt:

The neutral uses one name on their firm site, another on LinkedIn, and a third on provider panel pages. Different middle initials, different formality, different professional suffixes. Each variation reads to a human as obviously the same person. To an AI engine, these variations create entity ambiguity, which suppresses retrieval confidence.

The neutral’s credentials are stated differently in different places. “Florida Supreme Court Certified Circuit Civil Mediator” on one page. “FL-certified mediator” on another. “Certified by Florida courts” on a third. The variations are factually equivalent. They are not retrievably equivalent.

The neutral’s specialties shift across sources. The firm page lists six practice areas. LinkedIn lists three. The AAA panel page lists two. The implicit signal to an AI engine is that the specialties are not stable claims, and an unstable claim is a discounted claim.

The neutral has stale information on one major source. Their firm bio was updated last year. Their LinkedIn was last updated four years ago. The bar association directory shows a phone number that has not been current since 2019. The engine reads the inconsistency as risk and reduces citation confidence accordingly.

Coherence is the most under-engineered of the four CITE conditions and the easiest to fix. The Foundation phase of the 90-day plan addresses it first.

I — Identification

Identification is the condition AI engines evaluate to decide what to retrieve for you. Coherence answers who this is. Identification answers what kinds of cases, what kinds of expertise, and what kinds of context.

The work here is specific. AI engines cannot retrieve generic claims. “Experienced commercial mediator” is, to a machine, a phrase with no informational content. “Has mediated 312 commercial contract disputes since 2009, including franchise terminations, shareholder disputes, and post-M&A indemnity claims” is a phrase rich with retrievable detail. The first cannot be cited against a specific query. The second can.

Identification has four practical sub-components.

Credentials must be presented with the issuing body, jurisdiction, year, and verification path. “Florida Supreme Court Certified Circuit Civil Mediator, certified 2011, registration number XXXX” is far more retrievable than “Florida certified mediator.” The specificity is not pedantry. It is the signal AI engines use to verify the claim against the certifying body’s roster.

Specialties must be claimed in a disciplined, non-overlapping form. A neutral who claims to handle “complex commercial, employment, family, construction, personal injury, and elder law disputes” reads to a human as accomplished and to an AI engine as unverifiable. Two to four well-defined specialties, each with case-type granularity, retrieve far better than six broad claims.

Caseload signals must be quantified. “Mediated several hundred employment cases” tells an engine almost nothing. “Has mediated 412 employment cases since 2014, with a settlement rate of 78%,” tells an engine everything. Numbers, dates, and ratios are the units of retrievable evidence.

Jurisdictional anchoring must be explicit. “Practices nationally” is not an anchor. “Based in Miami; admitted to the Florida Bar, the Southern District of Florida, and the Eleventh Circuit; available for in-person mediations in South Florida and remote mediations nationwide” is. 

The geographic specificity is what allows the engine to surface you against city-specific and jurisdiction-specific queries.

T — Triangulation

Triangulation is the condition AI engines evaluate to decide whether to trust the claims you make about yourself. Self-attestation alone is not enough. The same claim, confirmed by three independent authoritative sources, becomes citable. 

The same claim made only by the neutral on their own site is discounted.

The triangulation logic is simple. AI engines are increasingly aggressively trained to detect and discount self-promotional content. They give weight to claims that can be cross-verified. A credential listed only on your firm's site is a self-claim. 

A credential listed on your firm site, on the certifying body’s roster, and on at least one third-party directory is a verified entity attribute.

For each central claim about your practice, ask: if someone wanted to verify it, where would they find independent confirmation? If the answer is “only on my own materials,” the claim is fragile. 

If the answer is “on at least three independent authoritative sources,” the claim is durable and citable.

The third-party sources that count are not all equal. The ADR Citation-Source Hierarchy below ranks them by the authority weight AI engines currently assign.

E — Evidence

Evidence is the condition that separates good neutrals from great ones in the AI retrieval game — and the condition almost no one in our profession has engineered.

Coherence, identification, and triangulation establish that you exist, that your credentials are real, and that your specialties are verifiable. Evidence establishes that you actually do the work at the level your credentials claim. 

Evidence is the content layer that demonstrates how you think, how you reason, how you approach the procedural and substantive challenges of your specialty.

Most neutrals have effectively zero evidence content. They have a bio. They have a list of credentials. They have a contact form. They have nothing — or close to nothing — that lets an AI engine, or an attorney reading what an engine returned, conclude that the neutral is excellent at their craft.

This is the Published-Reasoning Doctrine, the most contrarian and most consequential position in this article: the most undervalued GEO asset for a mediator or arbitrator is the published expression of how they reason. 

Not testimonials. Not case results, which most neutrals cannot disclose. Not generic blog posts. Reasoned content — procedural commentary, methodology explanation, framework analysis, hypothetical case reasoning, and published award commentary where permissible.

Evidence is what makes a neutral citable on their merits, not merely on their credentials. The fourth phase of the 90-day plan is built around producing it.

The ADR Citation-Source Hierarchy

The triangulation condition in CITE depends on understanding which third-party sources AI engines actually weigh as authoritative for neutrals. 

This is the second place where generic GEO advice fails our profession. The authority hierarchy for ADR differs from that for legal services generally, and the differences matter.

The six tiers, from the highest authority signal to the lowest:

Tier 1 — Institutional and Governmental Sources. Court-rostered neutral lists, state supreme court mediator certification databases, federal district court mediator panels, agency neutral rosters (EEOC, FMCS, NLRB). These sources carry the highest authority weight in AI retrieval because they are government-domain, verifiable, and not commercially manipulable. 

A neutral confirmed on a state supreme court certification database is, in AI engine terms, gold-tier verified.

Tier 2 — Provider Rosters and Bar Institutions. AAA, JAMS, FedArb, NAM, CPR, CEDR, and IBA panels. ABA Section of Dispute Resolution rosters and publications. State bar ADR section directories. These confer institutional authority on the neutral as an entity. 

The structural constraint discussed in Anomaly 2 applies — the panel page itself is often limited as a retrieval surface — but the fact of panel membership is a powerful verification signal that AI engines weigh heavily.

Tier 3 — Specialty Directories and Ranking Publications. Mediate.com, Lawsuit.com (mediator directory), Martindale, Chambers, Best Lawyers, Lawdragon, Super Lawyers, Who’s Who Legal. AI engines treat these as third-party verification sources, but they have become more discriminating about which directories they consider authoritative. 

Publications with editorial selection processes (Chambers, Lawdragon) currently carry more weight than publications with peer-vote or paid-placement components.

Tier 4 — Bylined Publications and Faculty Appearances. Bar journal articles authored by the neutral, Dispute Resolution Magazine and equivalent peer-reviewed ADR publications, CLE faculty pages, conference programs (ABA SDR, AAA-ICDR Foundation, NAFCM, ACR). This is the highest-yield earned-media tier that an individual neutral can actively control. 

Tier 1 and Tier 2 sources are largely earned through credential acquisition and panel application. Tier 4 sources are earned by producing and publishing work, making them the most direct mechanism by which a neutral can engineer additional authority.

Tier 5 — Owned and Social Sources. Personal practice websites, firm websites, LinkedIn profiles, Google Business Profiles, YouTube channels, podcast appearances, Avvo, Justia, Martindale (non-rated listings). Necessary but base-camp. AI engines weight these significantly lower than Tier 1–4 sources because they are self-attestation surfaces. 

The tactical implication is that owned and social sources are foundations, not summits — they must be coherent and identified, but they cannot carry the authority load alone.

Tier 6 — Published Reasoning. Published arbitral awards (with reasoning) in commercial cases, reported decisions citing the neutral’s process, published settlements where confidentiality permits attribution, methodology commentary on how the neutral approaches their work. 

AI engines treat this content category as evidence of competence rather than mere existence. Very few neutrals actively cultivate this footprint. The neutrals who do are dramatically over-cited relative to their objective credentials, which is the most exploitable opportunity in ADR-GEO today.

The Platform-Mediated Authority Reality

Tier 2 deserves additional honest discussion, because it is where most neutrals encounter the structural constraint described in Anomaly 2. Neutrals on rosters like AAA, JAMS, FedArb, NAM, and CPR benefit from significant institutional authority. The fact of panel membership is, by itself, a strong verification signal that AI engines recognize and weigh.

But the panel bio page is a different question from the panel membership. Most provider-platform bio pages are constrained by design — with limited outbound links, schema customization, content depth, and standardized formatting. 

This is not a criticism. The curation discipline that makes a provider roster valuable in the first place is the same discipline that limits how much an individual neutral can customize their representation on it. The constraints exist for legitimate reasons: fairness across panelists, brand consistency, and operational scale.

The practical consequence is straightforward: roster membership is necessary for top-tier authority signaling, but it is radically insufficient as a complete retrieval strategy. 

The panel bio handles institutional verification. It does not, by itself, build the coherent, identified, triangulated, and evidence-rich entity profile that AI engines need to cite a neutral confidently.

The answer is not to leave the panels. The answer is to build owned-authority surfaces in addition to panel membership — surfaces where the neutral has design and content control that the panel page cannot offer. The 90-day plan that follows is the systematic approach to doing exactly that.

The 90-Day Plan: Days 1–30 — Foundation

The remaining sections of this article translate the CITE framework into a concrete, sequenced execution plan. 90 days is the realistic window for a working neutral to move from the initial GEO assessment to an operational citation footprint. The plan is structured in three thirty-day phases.

Days 1–30 build the Foundation: the Coherence and Identification work in CITE, executed across the surfaces you already own or substantially control. Days 31–60 build Authority: the Triangulation work, executed across third-party sources. 

Days 61–90 build Amplification: the Evidence work, plus the measurement systems that tell you whether the previous sixty days have moved your citation footprint.

Foundation is unglamorous. It is also the phase when the largest gains occur with the least effort, because most neutrals have done so little that even basic execution produces visible movement in AI retrieval within weeks.

Week 1 — Entity Coherence Audit

The first week is diagnostic. Before fixing anything, document where you stand.

Open a spreadsheet. List every authoritative source where you appear publicly: your firm site bio, your LinkedIn, your AAA or JAMS panel page, every court roster, every bar directory, every specialty directory (Mediate.com, Martindale, Chambers, etc.), your Google Business Profile, your YouTube or podcast channels if any, and any bylined articles or published faculty pages you can find.

For each source, record exactly seven fields:

• The exact name as displayed

• The exact professional title and credential string

• The jurisdictions claimed

• The specialties claimed

• The contact information shown

• The headshot used (or “none”)

• The last-updated date if visible

Most neutrals doing this exercise for the first time produce a spreadsheet that surprises them. Names vary. Credentials are stated differently. Specialties have drifted across sources over the years as the practice has evolved. Three contact phone numbers appear across five listings. The headshot from 2018 is still on the bar directory.

This spreadsheet is the baseline. Everything in Weeks 2 and 3 is the work of bringing the rows into agreement.

Week 2 — Canonical Identity Definition

Before correcting any of the inconsistent sources, define the canonical version of your identity. Decisions to lock down, in writing:

The exact name you will use professionally on every source. Pick one form and use it everywhere. If you commonly go by “Bob” but your formal credentials are issued to “Robert J. Levin,” you have to pick — and the pick should be the form that matches your most authoritative credentials. Inconsistency between formal and informal forms across sources is one of the most common coherence failures.

The exact credential string for each credential you hold. Write out the canonical statement: “Florida Supreme Court Certified Circuit Civil Mediator, certified 2011, registration number XXXX.” Every source that mentions this credential should use this exact phrasing or a clearly compatible shortened form.

The two to four specialties you will claim. Be disciplined. A neutral with too many specialties reads as unverifiable. A neutral with too few may underclaim their actual practice. The right answer for most working neutrals is two to four well-defined specialty areas, each with three to six case-type sub-categories listed underneath.

The jurisdictional anchoring statement. Where you are based, where you are admitted, where you practice in-person, and where you practice remotely. Five or six lines of structured statement, not a vague phrase.

The current headshot. Choose one. Use it everywhere. A neutral whose headshot varies across major sources fragments their entity profile in a way that is easy to fix and consequential to leave unfixed.

The contact information format. One office address (or the explicit “available remotely; office address available on request” phrasing if you are virtual-only), one phone number, one professional email, and one website URL.

When this is written down — in a single document, no longer than a page — you have a canonical identity. The remaining work is propagation.

Week 3 — Propagation Across Surfaces You Control

Update, in order:

Your firm or practice website bio. This is your most-controlled surface. The canonical identity goes here first, in full, with every detail. This becomes the master version that other sources will reference.

Your LinkedIn profile. Align the name, title, credentials, specialties, jurisdiction, contact format, and headshot exactly to the canonical version. The summary section should be a condensed version of your firm bio, not a different bio with different claims.

Your Google Business Profile. The category, service area, description, photos, and credentials. GBP is enormously underweighted by most neutrals as a GEO surface, and it is one of the highest-leverage corrections in the Foundation phase.

Your directory listings on Avvo, Justia, Martindale (non-rated), and any other directories where you have profile control. Each one was updated to match the canonical version.

Your social profiles, where they exist publicly — author bios on any platforms where you publish, speaker bios from past conference appearances, if those pages are still indexed.

Week 4 — Provider Panel and Roster Reconciliation

The final week of the Foundation phase addresses the surfaces where you have limited control: provider panels, court rosters, and bar directories.

For each one, identify whether the bio shown is current, accurate, and aligned to your canonical identity. Where it is not, request an update through the provider’s standard channel. Provider rosters generally accept neutral-requested updates for factual content — credentials, contact information, specialty descriptions, headshot — even when the platform’s design constraints limit what you can change about the bio’s structure.

For the sources where requested updates take time to process (some provider platforms can take 60–90 days to push edits live), submit the request now so the change is live before the end of the 90-day plan.

Where a roster page genuinely cannot be brought into alignment with your canonical identity — because the platform’s design does not permit it — note the gap. The Authority phase, in Days 31–60, will build owned surfaces to compensate for the constraint.

At the end of Day 30, your coherence is engineered. The same name, the same credentials, the same specialties, the same headshot, the same contact information now appear consistently across every surface you control or can request updates from. 

Your identification is structured. Credentials are presented with verifiable specificity. Specialties are disciplined. Jurisdictional anchoring is explicit.

You have not yet built any new third-party authority — that is the work of Days 31–60. But you have done something most neutrals never do: you have made yourself coherent and identified across the web. 

That alone, in our observation of AI retrieval behavior, is enough to produce visible movement in citation patterns by Day 30. The Authority phase that follows substantially compounds the movement.

The 90-Day Plan: Days 31–60 — Authority

The Foundation phase made you coherent and identified. The Authority phase makes you triangulate.

This is the work where most neutrals stall. Foundation can be executed unilaterally — you control the surfaces, you make the edits, and the work is done. 

Authority requires reaching beyond your own surfaces to third-party sources that provide verification. This work is slower, less controllable, and uncomfortable for neutrals who have spent careers letting their reputation speak for itself.

The discomfort is misplaced. Authority is not self-promotion. It is the systematic establishment of independent third-party confirmation for claims you are already making about your practice. 

The work is not asking the world to believe you are excellent. It is making sure that when an AI engine — or an attorney, or a party — checks whether your claims are verifiable, the verification is present.

Week 5 — Tier 1 and Tier 2 Audit

The first week of the Authority phase targets the highest-weight authority surfaces in the citation-source hierarchy: institutional and governmental sources and provider rosters.

For Tier 1 sources, the question is whether you appear, completely and accurately, on every governmental or court-affiliated roster you are eligible for. Most neutrals appear on some but not all. Common gaps:

• State supreme court certification databases that are publicly searchable, but where the neutral’s listing is incomplete or out of date

• Federal district court mediator panels in jurisdictions where the neutral practices but has not formally registered

• Agency-neutral rosters (EEOC, FMCS, NLRB, state human rights commissions) that admit external neutrals on application

• State bar ADR section directories that require active section membership but are otherwise free to join

Identify every Tier 1 source for which you are eligible. Where you are listed, confirm the listing is current and accurate. Where you are eligible but not listed, begin the application or registration process. Some applications take weeks to process; starting them in Week 5 means many will be live by Day 90.

For Tier 2 sources, the question is different. Provider panel applications take months and have selectivity. AAA, JAMS, FedArb, NAM, and CPR are not memberships you join in a 90-day window. But there are Tier 2 sources that are accessible within the window:

• Your state bar ADR section directory and any specialty bar associations

• The ABA Section of Dispute Resolution, which is a paid membership but immediately confers Tier 2 directory presence

• Specialty-area ADR organizations (the Council for Construction Mediators, the Employment Law Alliance neutrals network, etc.), where membership is application-based but processable within 30–60 days

The work in Week 5 is mapping every Tier 1 and Tier 2 source you should be on, comparing it to where you currently are, and beginning to close the gaps that can be addressed within the 90-day window.

Week 6 — Specialty Directory Optimization

Week 6 addresses the Tier 3 layer: specialty directories and ranking publications.

For each Tier 3 directory you appear on — Mediate.com, Lawsuit.com, Martindale, Chambers, Best Lawyers, Lawdragon, Super Lawyers — the work is the same: bring the listing into alignment with your canonical identity from Week 2, and use every available field to maximum density.

Specialty directories are unusual in the GEO landscape because they offer more bio control than provider panels but less than your own website. 

The fields available — practice areas, jurisdiction, experience level, credentials, biographical narrative, articles, case results where permissible — are largely under your control once you are listed. Most neutrals fill these fields once at signup and never return. 

The neutrals who maintain their specialty directory listings as actively as they maintain their LinkedIn profiles are dramatically overrepresented in AI retrieval.

The work in Week 6 is straightforward but tedious: for each Tier 3 directory, log in, expand every available field to its maximum density using your canonical identity, refresh the headshot if needed, and ensure the credential and specialty statements match your canonical version exactly.

A specific note on directories that allow articles or publications to be attached to the profile: this feature is enormously underused. 

A specialty directory profile with three or four attached articles is treated by AI engines as substantially more credible than a profile with a biographical narrative alone. 

The Amplification phase will produce content; the directory listings should be updated to attach that content as it appears.

Week 7 — Bylined Publication Pitch

Week 7 begins the most consequential Authority work: securing a Tier 4 publication.

Tier 4 sources — bar journal articles, Dispute Resolution Magazine, CLE faculty appearances, conference programs — are the highest-yield earned-media tier that an individual neutral can actively control. 

They are also the work most neutrals avoid because the path is slower than other Authority work, and the rejection risk is real.

The pitch process is straightforward in structure, even if individual outcomes are uncertain. The recommended approach for Week 7:

Identify three publications appropriate to your specialty and seniority. For most neutrals, this means one bar journal (state bar magazine or specialty bar publication), one ADR-specific publication (Dispute Resolution Magazine, Alternatives from CPR, or a similar journal), and one practitioner-facing publication adjacent to your specialty (a construction industry magazine, an employment law journal, a family law newsletter).

For each, identify the editor or submissions process. Most legal and ADR publications have public submission guidelines.

Draft three pitch concepts. Each should be a 200-word abstract proposing an article you could write. The strongest pitches are not about your practice — they are about a specific procedural, substantive, or methodological question in your specialty area where you have a defensible point of view. “Why early case evaluation fails in construction defect mediations” is a stronger pitch than “Best practices in mediation.”

Send the three pitches in Week 7. The realistic landing rate is approximately 1 in 3 for cold submissions to mid-tier publications, somewhat lower for top-tier publications. Plan to repeat the process in Week 9 if all three Week 7 pitches are declined.

Critically, an accepted pitch in Week 7 will not produce a published article by Day 90. Publication timelines for legal and ADR journals are typically three to six months. 

The Authority work in Week 7 lays the foundation for citations that will go live in months four through seven of your GEO build, not within the 90-day window itself. The 90-day plan is the starting horizon, not the complete horizon, for Authority work.

Week 8 — CLE Faculty and Conference Speaker Outreach

The final week of the Authority phase targets the speaking and faculty pathways into Tier 4.

CLE programs in every major jurisdiction accept faculty proposals from practicing neutrals. The state bar’s CLE provider, the ABA, specialty bar associations, and private CLE providers all run regular programming and need expert faculty.

The path is similar to the bylined publication path: identify the provider, identify their submission process, propose a 60- to 90-minute program in your specialty, and submit.

Conference speaking is similar. Annual ADR conferences — the ABA Section of Dispute Resolution Spring Conference, AAA-ICDR Foundation events, regional ACR conferences, specialty conferences in construction or employment ADR — all run call-for-proposals processes. 

The proposal deadlines for major 2027 conferences typically open in mid-2026 and close by fall, which means Week 8 is the right moment to be submitting proposals for events that will publish your speaker bio by year-end and produce live conference appearances in the first half of 2027.

The compounding effect of CLE faculty and conference speaker appearances is significant. Each program produces, at minimum, a speaker bio page on a Tier 4 source — itself a citation-grade authority signal. 

Many programs produce additional artifacts: recorded sessions on the provider’s platform, course materials distributed to attendees, post-conference write-ups in trade publications, and social-media coverage from attendees. A single accepted CLE proposal can generate four to eight Tier 4 citations across the eighteen months following acceptance.

The Authority phase, executed as described, will not produce visible citation outcomes by Day 60. The work is structurally slower than the Foundation work because it depends on third parties to accept, schedule, and publish. 

What the Authority phase produces by Day 60 is the pipeline — the submitted pitches, the active applications, the proposed CLE programs, the directory profiles at maximum density — that will yield Tier 3 and Tier 4 citations across the second half of 2026 and into 2027.

The 90-Day Plan: Days 61–90 — Amplification

The Amplification phase brings the work back under your direct control after the Authority phase’s necessary dependencies on third parties. The work here is what closes the loop on CITE: the Evidence condition, executed through the Published-Reasoning Doctrine, structured around an AI-readable bio template, and instrumented with the measurement systems that tell you whether the previous sixty days are working.

Week 9 — The AI-Readable Bio Build

The bio you have been polishing since Week 1 is, in most cases, written for human readers. Even when it is detailed, current, and accurate, it is structured as professional bios have been for decades: prose paragraphs, narrative flow, achievements woven into sentences, credentials introduced in context.

This structure is suboptimal for AI retrieval. AI engines parse content in passages, extract claims as discrete facts, and reassemble those facts into responses. A bio written as flowing prose is harder for an engine to parse cleanly than a bio written with structural discipline.

Week 9 is about rebuilding your bio for AI readability — without sacrificing human readability. The two are compatible; most neutrals just have not done the structural work.

The structure of an AI-readable neutral bio, in order:

Identity block. Full name, professional designations, and location of practice. One line. Specific. “Robert J. Levin, Esq. — Mediator and Arbitrator, Aventura, Florida.”

Headline statement. One sentence that positions the neutral specifically enough to be retrievable against a real query. The headline answers: what kind of cases, in what kind of practice, at what scale. 

Not “experienced mediator and arbitrator” but “Florida-certified circuit civil mediator focused on commercial contract and employment disputes, with 312 mediations conducted since 2014.”

Credentials section. Structured list, not prose. Each credential gets its own line, with issuing body, jurisdiction, year, and registration number where applicable. The structure matters: AI engines extract structured lists more reliably than embedded prose mentions.

Practice areas. Two to four specialties, each named clearly. Under each specialty, three to six case-type subcategories are listed in a structured format.

Caseload signal. Quantified statement: number of cases, time period, settlement rate, or comparable outcome metric where appropriate. This is the single highest-leverage paragraph in the bio for retrievability.

Procedural orientation. One paragraph — three to five sentences — describing how the neutral approaches their work. This is the AI-readability paragraph that allows engines to retrieve the neutral for “what kind of mediator is X” queries. It is the place where personality, methodology, and approach get expressed in a form a machine can parse.

Notable appointments and roles. Court rosters, provider panels, bar leadership, faculty roles, board service. Structured list, with the institutional name and the neutral’s role on each line.

Publications and presentations. Recent first; last 18 months weighted most heavily. Each entry includes title, publication or venue, and date.

Verification footer. Direct links to authoritative third-party sources that confirm the central credentials. Not a generic “see my credentials at...” but specific deep links to the certifying body’s roster entry, the major panel pages, the bar directory.

Alongside this human-readable bio, the page should include JSON-LD structured data — invisible to human readers but directly readable by AI engines. The schema marks up the same information in machine-readable form, eliminating any ambiguity about which credentials, specialties, and affiliations belong to the neutral.

The full JSON-LD example, with annotated fields, is included in the canonical asset embedded in this section. The schema combines the Person, Attorney, and LegalService types, along with credentialing relationships, jurisdictional attributes, and organizational affiliations. 

A web developer can implement the schema in under an hour, given the template. Most neutrals’ sites currently have no person-level structured data at all.

Week 10 — The First Piece of Published Reasoning

The Published-Reasoning Doctrine, introduced earlier in this article, is the most important strategic move available to most working neutrals — and the move almost none of them are making.

The thesis bears restating in its strongest form: the most undervalued GEO asset for a mediator or arbitrator is not directory presence, not testimonials, not credential polish, not even publication frequency. 

It is the published expression of how the neutral reasons. Procedural commentary. Methodology explanation. Framework analysis. Hypothetical case reasoning. Commentary on published awards or reported decisions where confidentiality permits.

The reason this content category is so undervalued is that most neutrals have concluded — correctly — that they cannot publish detailed accounts of cases they have handled. 

From that correct conclusion, they have drawn the incorrect inference that they cannot publish content that demonstrates expertise.

Confidentiality does not prevent published reasoning. It forces a better kind of published reasoning. The constraint generates content discipline that, when done well, produces material that AI engines weigh more heavily than case studies would.

Week 10 is the work of producing your first piece of published reasoning. The realistic target is a single piece, 1,500 to 2,500 words, published on your own site, structured for both human readability and machine retrieval.

The topic should be in your specialty. The piece should answer a real question that working attorneys, parties, or other neutrals would search for. Strong examples:

• “What kind of mediator do you want for a partnership dissolution? Five factors that change the answer.”

• “How I approach the first hour of a complex commercial mediation — and why most mediators handle it wrong.”

• “The procedural mistake I see most often in employment arbitration demand letters.”

• “When early case evaluation fails: three diagnostic signs construction defect cases will not settle pre-discovery.”

Each of these is a topic on which a working neutral has, by definition, a defensible point of view. Each can be written without violating confidentiality — none require disclosure of cases, parties, or outcomes. Each answers a real question that real users type into AI engines. 

Each, once published, becomes a Tier 6 source that an AI engine can cite when answering related queries.

The structural requirements for the piece, beyond the topic:

A clear question is stated in the first paragraph. AI engines parse content as question-answer pairs more than as narrative flow.

H2 and H3 headings that mirror the question's substructure. “Five factors” should produce five H3-level sections, each named after the factor.

Specific examples, hypotheticals, or numerically anchored claims. Generic discussion is retrievable; specific claims are citable.

A “key takeaway” paragraph at the end of each major section, suitable for direct extraction by an AI engine. This is the GEO version of the featured-snippet optimization technique from traditional SEO, applied to AI passage retrieval.

A bio block at the end of the piece, in the AI-readable structure described above.

One piece in Week 10. Subsequent weeks of subsequent quarters add more. The compounding effect of consistent published reasoning is significant: by the end of the first year, a neutral who publishes one substantive piece per month has built a 12-piece content footprint that AI engines treat as a defensible signal of demonstrated expertise.

Week 11 — Measurement Setup

Most GEO strategies fail at measurement. Not because measurement is technically difficult, but because most neutrals never set it up, run it once, and never repeat it.

The minimum viable measurement system for an individual neutral consists of three components.

Component 1 — The recurring prompt set. A standardized set of 10 to 15 queries, run monthly across the same six AI engines used in the baseline test at the beginning of this article. The queries should include the geographic and specialty queries most relevant to your practice, plus a small number of more general queries where your specialty is relevant. 

The same queries run every month from the same clean browser environment and are recorded in the same spreadsheet. The point is not absolute accuracy on any single run; the point is detecting change across months.

Component 2 — The citation log. A running document where every confirmed AI citation of your name or practice is recorded with date, engine, query that produced it, and screenshot. Citations that appeared in Month 1 but disappeared in Month 2 are recorded as recovered or lost. 

The log is a longitudinal record of citation footprint and serves as the data backbone for understanding what is actually working.

Component 3 — The third-party authority log. A list of every Tier 1 through Tier 4 source on which your listing has been updated, applied for, or earned during the 90-day plan. Each entry includes the source, the date of submission or update, and the status. This log connects effort to outcome — when a Tier 4 publication accepts your article in Month 4 of 2027, the log shows the dated through-line from the pitch in Week 7 of the original 90-day plan.

Week 12 — The Forward Plan

The final week of the 90-day plan is about looking past Day 90.

By Day 90, the Foundation work is complete, and the citation footprint is visibly different from the baseline. The Authority work is in the pipeline — applications submitted, pitches sent, programs proposed, with delivery dates running through the following twelve months. 

The Amplification work has produced its first piece of published reasoning, and the measurement systems are running.

What comes next is the operational rhythm that turns this 90-day plan into an ongoing practice. The recommended rhythm:

• One piece of published reasoning per month, minimum

• One new Tier 4 pitch or proposal per quarter

• One quarterly review of the recurring prompt set and citation log

• An annual full re-audit of the entity coherence baseline (because circumstances change, credentials evolve, and even well-maintained coherence drifts over twelve months)

At Mediate Lawsuit, we are building infrastructure to support this rhythm at scale. The forthcoming CITE Score — launching alongside The State of Mediation in Q4 2026 — will provide individual neutrals with a structured four-axis assessment of their citation readiness across Coherence, Identification, Triangulation, and Evidence. Early access is available to neutrals who join the waitlist at MediateLawsuit.com/cite-score.

The State of Mediation: A Benchmark for the Profession

This article is the first in a connected body of work Mediate Lawsuit is launching to make the digital visibility of the ADR profession measurable, comparable, and improvable.

In Q4 2026, we will publish the first annual edition of The State of Mediation: a benchmark report on digital visibility patterns across the mediation and arbitration profession in the United States. The report will measure, by geography, specialty, and provider affiliation:

• The distribution of AI citation frequency across active neutrals

• The Tier 1 through Tier 6 authority footprint patterns that correlate with citation

• The earned-media activity patterns of the most-cited neutrals in each specialty

• The growth in AI-mediated discovery as a share of the total selection-layer market for neutrals

• The gaps between objectively credentialed neutrals and actually-cited neutrals — and the structural reasons those gaps exist

The methodology will combine the prompt-test architecture described in this article, scaled to a representative sample of specialties and geographies, with the third-party source audit methodology applied to a representative sample of provider panel surfaces.

The report is built to be useful to three audiences. Working neutrals will use it to benchmark their own citation footprint against peer ranges. 

ADR providers will use it to understand the visibility landscape their panels operate. Researchers and educators in the dispute resolution field will use it as a longitudinal record of how the selection layer is changing.

Early access to the first annual edition, including a preview of methodology and a notification when the report is published, is available at MediateLawsuit.com/state-of-mediation.

What Happens Next

The choice in front of every working mediator and arbitrator is no longer whether to engage with AI-mediated discovery. AI engines are already mediating the discovery process for an increasing share of the attorneys and parties who need neutrals. The choice is whether to be engineered into the answers those engines return, or to be absent from them.

The CITE framework — Coherence, Identification, Triangulation, Evidence — defines whether a neutral is citable. The 90-day plan turns that structure into execution. 

The ADR Citation-Source Hierarchy and the AI-Readable Neutral Bio Template provide the reference assets. The forthcoming State of Mediation benchmark will provide the ongoing record of how the profession’s visibility is shifting.

What is left is the work.

Most neutrals will not do it. The profession’s traditions — reputation through word-of-mouth referral, authority through credentials accumulated over decades, dignity that resists self-promotion — will continue to lead many practitioners to assume that the changing selection layer is someone else’s problem. They will be wrong, and the cost of being wrong will compound.

The neutrals who execute the work in this article — even imperfectly, even partially — will be measurably more visible to the AI engines that increasingly drive neutral selection. The compounding advantage of that visibility, over the next two to five years, will be substantial.

Get Started

Claim your enhanced Mediate Lawsuit directory profile. Mediate Lawsuit’s directory is architected around the CITE framework. Every neutral profile is structured for AI retrieval, with the coherence, identification, triangulation, and evidence layers built into the platform from the ground up. 

Enhanced profiles include the full AI-readable bio template, JSON-LD schema integration, published-reasoning hosting, and verification anchors to your Tier 1 and Tier 2 authority sources. Join the directory at MediateLawsuit.com/join.

Join the CITE Score waitlist. The free CITE Score audit launches in Q4 2026 alongside The State of Mediation. Waitlist members receive early access and a preview report on their current citation readiness. Sign up at MediateLawsuit.com/cite-score.

Read the forthcoming State of Mediation benchmark. The first annual report publishes Q4 2026. Receive notification and a methodology preview at MediateLawsuit.com/state-of-mediation.

For practices that need help implementing the CITE framework end-to-end, Transformational Outsourcing’s 90-day GEO program builds the full stack on the neutral’s behalf — entity coherence engineering, schema implementation, earned-media pipeline development, and published-reasoning production. Inquiries at TransformationalOutsourcing.com.

Robert (Bob) Levin is CTO of Mediate Lawsuit and Founder of Transformational Outsourcing. He has built digital infrastructure for ADR practitioners and legal services firms since 2008, and is the architect of the CITE framework and the forthcoming State of Mediation benchmark.

This article is a living document. Updates and revisions are tracked openly. Last updated: [06/01/2026].