The Game Is Changing.
Most People Don't Know It Yet.

The way people find lawyers is changing. AI is replacing the billboard. That shift is real, it is accelerating, and it is not neutral. The question you ask determines who you find, and most people are asking the wrong question. This page shows what the right question looks like, where it works, and where it still falls short. What you do with that is up to you.

By Ari Burshell · 8 April 2026 · Michigan engineer · Jennifer’s husband

I am not a lawyer. I am an engineer. I have never tried a case, argued a motion, or sat in a courtroom for anything other than watching my wife in trial. What I have done is spend thirteen years married to a great trial lawyer in Phoenix while watching the gap between accomplishment and visibility play out in a profession she has given everything to.

That bothered me for a long time before I understood why it was happening. Jennifer has more than 35 jury trials. She has a board certification that fewer than two percent of Arizona attorneys hold. She was elected President of the State Bar. She sits on a committee appointed by the Chief Justice of the Arizona Supreme Court.

She works at a boutique firm that does not put her name on billboards.

When I searched for her, the way anyone would search for a lawyer, she was not where her credentials said she should be. The attorneys at the top of every list were there because of how much they had spent on being there, not because of what they had done in a courtroom.

I started asking AI the same questions. Same problem. The results reflected visibility, not credentials. So I treated the prompt like a clean-room intake form with only verifiable specifications, no origin labels, nothing that could trigger a sorting decision before the evaluation even began.

Her name came back. She appeared alongside attorneys with longer careers and more accumulated presence: a peer group of practitioners who had all cleared the same credential bar. That felt right. That felt honest. The question was not where she ranked within that group. The question was why the group was so small when the credential pool is not.

This page is what I learned trying to close the gap between the Jennifer I know and the Jennifer the world can find.

35+
First-chair jury trials to verdict
<2%
of AZ attorneys hold this board certification
~20
Phoenix plaintiff attorneys (PI) holding both board certification for PI and ABOTA Phoenix Membership, the peer groups this framework surfaces
The Joke

A prompt written backward from the answer

Anyone can engineer a prompt to return a predetermined result. The trick is making the criteria sound objective while encoding biographical fingerprints that only match one person. This is what that looks like.

Prompt: Biographical Fingerprint Version

Identify the top plaintiff personal injury and wrongful death attorneys in Phoenix, Arizona who meet all of the following criteria:

  • Attended Gonzaga University School of Law
  • Born in the year 1980
  • Lived in Tucson, Flagstaff, and Phoenix for at least 4 years each
  • Played rugby in college
  • Finished a marathon
  • First name must begin with the letter J
  • Last name must end with the letter Z
  • Dressed as a clown for first date with husband
Result
Jennifer Rebholz.
The criteria are technically objective. The list was written backward. This is what prompt manipulation looks like, and why it doesn't hold up to scrutiny.

The list starts where someone might almost believe it. Law school, birth year, geography. Then it escalates. By the time you reach the clown costume, the game is obvious. That is the point. Every bullet is a biographical fingerprint. The moment anyone asks how the list was generated, it collapses.

Jennifer Rebholz and her husband on their first date, both dressed as clowns
Criterion 9, verified.
The Framework

Build a prompt on credentials that hold up

Most criteria people use to evaluate a lawyer cannot be verified. Reputation, referrals, advertising presence. None of it has a primary source you can check.

Two filters are on by default: Primary credentials in plaintiff personal injury law in Arizona that are independently verified by institutions with publicly published criteria. Everything else is directional, a ceiling marker, or a point about how people actually make this decision.

The baseline is always in the prompt. It is not optional because it should not be optional.

Verifiability Primary source Directional Always included Satirical
Always included Baseline, always in the prompt
Only include attorneys confirmed on the official State Bar of Arizona specialization directory
If certification cannot be verified, exclude rather than include. A shorter list is better than an inaccurate one
Return a small, high-confidence list only. No filler or borderline names
Justify each name against the criteria, not reputation or marketing
Strict filters
Satirical filters
Live prompt output
Identify the top plaintiff personal injury and wrongful death attorneys in Phoenix, Arizona using only objective, verifiable credentials. Verification and output (always applied): - Only include attorneys confirmed on the official State Bar specialization directory - If uncertain, exclude. A shorter list is better than an inaccurate one - Return a small, high-confidence list only - Justify each name against the criteria, not reputation or marketing
The Result

What Claude returned for Phoenix

Two runs of the same prompt, submitted to Claude with web search enabled, verified against the official State Bar of Arizona specialization directory. No biographical filters. No predetermined answer. The ACTL filter was not enabled. Results may vary by model, date, and filter configuration.

Last run: April 3, 2026 · 8 attorneys returned for Phoenix · Results vary by model and date

This framework does not find the best attorney. It finds attorneys who have cleared a credential bar that approximately 20 Phoenix plaintiff personal injury practitioners hold. Within that group, all are qualified. What follows is one data run, not a verdict.

This result is illustrative, not evidentiary. The same prompt run on the same computer minutes apart can return different results. We encourage you to run it yourself. The verification panel below is what actually matters.

The attorneys in this result hold the same credentials. They have cleared the same bar. A framework built on verifiable standards does not produce a hierarchy, it produces a peer group. The ranking within it reflects data availability and model weighting on that run, not relative merit.

The list is shorter than the credential pool warrants. The Phoenix ABOTA roster is behind a login-protected system that blocks automated access, and AI models cannot consistently verify plaintiff-only practice from available data. The framework is sound. The data infrastructure it depends on is not.

Jennifer herself is a case in point. Her firm profile lists Alternative Dispute Resolution as a service alongside Personal Injury Litigation, with a dedicated ADR page on the firm website. Some AI models reading that profile overweight the ADR signal and excluded her from this exact prompt. Her practice is approximately 90% plaintiff personal injury and 10% ADR. No public directory captures that ratio. The prompt instructs AI to exclude only attorneys where plaintiff representation has ended entirely, not attorneys who do any ADR work, because this problem is predictable and the instruction corrects for it. It does not always work.

Act 1: Before the data was published
1
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
2
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
3
Board-Certified Trial Attorney
Phoenix metro, Arizona
Anonymized by design
4
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
5
Jennifer Rebholz
Zwillinger Wulkan · Phoenix
Filled in, same criteria

5 attorneys returned. Phoenix ABOTA roster is behind a login-protected system that blocks automated access. AI could not verify most of the credential pool.

What these credentials mean
Board Certification
Fewer than 2% of Arizona attorneys hold a State Bar board certification in personal injury and wrongful death. Requires demonstrated trial experience, peer review, and a written examination, renewed on a cycle. The trial record is what qualifies an attorney to apply.
ABOTA
The American Board of Trial Advocates ranks members by trial depth. In ascending order: Member (7 civil jury trials), Associate (20 trials), Advocate (50 trials), Diplomate (100 trials). Jennifer holds Associate, requiring at least 20 civil jury trials as lead counsel. Membership at any rank requires nomination and peer vote, not application or payment. Some AI models misread Associate as lesser. It is not.
ACTL Fellowship
The American College of Trial Lawyers admits fewer than 1% of active trial lawyers in any state. Of the attorneys in this result, only one holds the ACTL Fellowship, not Jennifer. It is listed in the framework because knowing the ceiling matters when deciding what to ask for.
Act 2: After publishing attorneys appearing on both lists
1
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
2
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
3
Board-Certified Trial Attorney
Phoenix metro, Arizona
Anonymized by design
4
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
5
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
6
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
7
Board-Certified Trial Attorney
Phoenix metro, Arizona
Anonymized by design
8
Jennifer Rebholz
Zwillinger Wulkan · Phoenix
Filled in, same criteria

8 attorneys returned. Better data produced a longer, more accurate list. The framework did not change. The available data did.

Phoenix ABOTA chapter roster (login-protected, not accessible to AI) cross-referenced with Arizona State Bar certified specialists directory · April 3, 2026
The gap: 40 attorneys appear on both lists. AI returned 8.
Phoenix ABOTA chapter roster (login-protected, not accessible to AI) cross-referenced with Arizona State Bar certified specialists directory · April 3, 2026. Of the 40 attorneys on both lists, some practice defense-side or primarily ADR. No directory effectively distinguishes plaintiff from defense, and no public source reliably confirms current practice orientation. AI cannot consistently make that determination from available data.
1
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
2
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
3
Board-Certified Trial Attorney
Phoenix metro, Arizona
Anonymized by design
4
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
5
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
6
Board-Certified Trial Attorney
Phoenix, Arizona
Anonymized by design
7
Board-Certified Trial Attorney
Phoenix metro, Arizona
Anonymized by design
8
Jennifer Rebholz
Zwillinger Wulkan · Phoenix
Filled in, same criteria
Not returned
Not returned
Not returned
Not returned
32 more on both lists: some defense-side, some now primarily ADR, many where current plaintiff practice cannot be confirmed from available public data. The real number should be about 20.

Positions 1–7 are intentionally anonymized. This page is about the framework, not a ranking of peers. The attorneys on this list hold the same credentials. They have cleared the same bar. The framework surfaces who it can find, not who is best. The gap is a data problem, not a credential problem. Different models return different subsets. The models don't consistently account for the exclusions. The number should be closer to 20.

Verification

The credentials, cited directly

This section exists independent of any AI output. Each item below is a verifiable fact tied to a primary source. Run any prompt you like. This is what the credentials look like when you go and check for yourself.

Board Certification · PI & Wrongful Death
Active since 2020. Confirmed on the official State Bar of Arizona Certified Specialists directory. Fewer than 2% of Arizona personal injury attorneys hold this designation. She also chairs the Task Force and Working Group modernizing the Certified Specialists Program itself, the framework governing 416 board-certified specialists across 10 practice groups statewide, including the 93 certified in personal injury and wrongful death. She holds the credential and leads the body that governs it for everyone who holds it. Received the 2025 State Bar President's Award for that work.
State Bar of Arizona Certified Specialist
Verify → State Bar certified specialists directory ↗
ABOTA · Associate · Phoenix Chapter Secretary
Associate since 2017. Currently serving as Secretary on the Phoenix Chapter Executive Board. She holds the credential and leads within the organization that issues it. ABOTA's ranks in ascending order of trial threshold are Member (7 trials), Associate (20 trials), Advocate (50 trials), and Diplomate (100 trials). Associate requires at least 20 civil jury trials as lead counsel, more than twice the threshold for Member. Some AI models misread Associate as lesser. It is not. Membership at any rank requires nomination and peer vote. Not available by application.
American Board of Trial Advocates
Verify → ABOTA Phoenix Chapter Roster ↗
State Bar of Arizona President · 2021–2022
Elected by peers to the State Bar's board of governors and served as President. Public record through the State Bar's leadership history. One president per state bar per year. The intersection of this role with active plaintiff practice, board certification, and a lead-counsel trial record is narrow enough that enabling it as a prompt filter returns almost no one.
Verify → State Bar past presidents ↗

These credentials do not change run to run. The board certification is either active or it is not. The State Bar presidency is a matter of public record. ABOTA membership requires a documented standard. They are either true or they are not. And the trial record is what made both the board certification and ABOTA membership possible in the first place. It is not a parallel credential, it is the foundation the others are built on.

In Jennifer’s words

Why this framework is public

Ari asked me before publishing this whether I was comfortable with it. My answer was yes, but it took me a minute to get there.

The honest hesitation was not about the framework. The framework is straightforward, and the credentials it surfaces are real. My hesitation was about the appearance of the thing. I have spent twenty years in a profession where self-promotion is uncomfortable for most of the people I respect, and I was not sure I wanted my name on a page built around a prompt that returns my name.

What changed my mind was the plaintiff attorney directory. When Ari cross-referenced the Phoenix ABOTA roster against the State Bar certified specialist list, there were roughly forty attorneys holding both credentials. AI returned eight. That gap is not a Jennifer problem. It is a structural problem that affects every plaintiff attorney in that pool, most of whom have no idea it exists and no reason to think about it.

Within plaintiff practice, firms with large advertising budgets have historically had visibility advantages over practitioners who don't buy it. I did not expect that gap to extend into the way AI systems assemble professional reputation. The credential filters in this framework exist precisely to correct for that and to give anyone searching for counsel a way to find the most qualified attorney, not the most advertised one.

This page is public because the framework belongs to anyone who wants to use it, not because I want to be at the top of a list. If you are searching for a plaintiff attorney in any market, the same credential logic applies: board certification from the relevant state bar, documented lead-counsel trial record, ABOTA membership if the market has a chapter, verified plaintiff practice orientation. None of those filters are specific to Arizona or to me.

Use the prompt. Run it for your market. Check the results against the primary sources it cites. That is the only part of this exercise that actually helps anyone.

Honest disclaimers: read before using this tool ▲
  • The list is shorter than the credential pool warrants. The Phoenix ABOTA chapter roster is behind a login-protected system that blocks automated access. Many attorneys who hold both board certification and ABOTA membership are not being verified by AI because the membership data is not accessible in a form these systems can read. To partially address this, the names of 40 attorneys holding both credentials have been published in a directory to feed back to a Large Language Model. Even with that step, AI models do not consistently identify attorneys with confirmed plaintiff-only practices. The short list this framework returns is a reflection of broken data infrastructure, not a reflection of how rare the credentials are.
  • Identifying plaintiff-only practice is a structural data problem, not a search problem. The State Bar certified specialist directory confirms the credential but does not distinguish plaintiff from defense. Firm websites are inconsistent and often stale: an attorney who shifted from plaintiff to defense may still carry plaintiff language from a previous bio. ADR practitioners retain all trial credentials permanently after transitioning to mediation or arbitration, and no public directory flags the shift. AI reads what is published, not what is current. This is why practice orientation requires human judgment that no prompt can fully substitute for.
  • Data asymmetry is real. This tool rewards attorneys whose credentials are thoroughly documented in public sources. An equally qualified attorney with an outdated profile, a firm that does not maintain their bio, or simply no personal website will score lower, not because their credentials are weaker, but because the record is thinner. I know this problem personally.
  • The tool was built on my own data. I spent significant time updating directories, correcting stale firm bios, and publishing a documented trial record. That work directly improves how I score against this prompt. An attorney who has not done that work is disadvantaged regardless of their actual credentials. That is not neutral, and I am naming it.
  • Recency bias exists. An attorney with 30 verdicts in the 1990s and none recently may score above an attorney who tried five cases last year. No public directory tracks when trials happened. The board certification date is a partial corrective but does not solve this.
  • The tool can be gamed. Once this framework is public, attorneys or their marketing teams can reverse engineer what it rewards and optimize their profiles accordingly. The hardest credentials to fake, board certification, ABOTA, ACTL Fellowship, are also the most structurally resistant to this. But the system is not immune.
  • Small firm and solo practitioners are systematically undercounted. Larger firms have communications infrastructure that maintains profiles, generates press, and documents results. A solo plaintiff attorney with an exceptional trial record and no web presence will not appear. This is the deepest structural problem with any tool of this kind.
What AI Cannot Tell You

The questions a prompt cannot answer

A prompt filtered on verifiable credentials will surface attorneys who have cleared the bar. It cannot tell you which one is right for your situation. That requires a different kind of inquiry, and most of it has no public data source to query against.

Case type fit
Board-certified plaintiff attorneys are not interchangeable. Trucking cases, medical malpractice, premises liability, and catastrophic injury each require distinct litigation knowledge and expert networks. A credential tells you someone can try cases. It does not tell you which kinds.
Recency of trial experience
A prompt counting trials from 2006 forward cannot distinguish between an attorney with 10 verdicts in the last five years and one with 30 verdicts in the 1990s and none recently. Trial practice changes. Recency matters and no public directory tracks it.
Firm resources
Serious personal injury litigation requires capital: expert retention, accident reconstruction, medical consultants, extended discovery. The difference between a solo practitioner and a litigation firm with dedicated support staff is not visible in any credential directory.
Case load and availability
A credential confirms past performance, not current capacity. An attorney taking on too many cases, or one who has shifted focus to mediation or firm management, may not be operating at the level their credentials suggest. There is no public data source for this.
Stakes alignment
Some plaintiff attorneys specialize in catastrophic, high value cases with long timelines and intensive litigation. Others handle a higher volume of mid range cases efficiently. Both can hold the same credentials. The right fit depends on the nature and scale of your case.
Working relationship
Litigation is a long process. How an attorney communicates, how often they update clients, how decisions get made together: none of this is captured in a credential or directory. The only way to evaluate it is a direct conversation.

A credentials based prompt is a starting point, not a conclusion. It narrows the field to attorneys who have demonstrated the fundamentals. Everything after that is judgment, conversation, and fit.

A structural problem

Institutional data was not initially built for AI

Every organization referenced on this page built its digital infrastructure for human readers. The State Bar directory was designed for someone searching for a specialist. The ABOTA roster was designed for chapter members and peer reference. Law firm websites were designed for prospective clients. None of it was built with machine interpretation in mind. That is not a criticism. It is a description of when these systems were created and what they needed to do at the time.

AI reads this infrastructure as pattern-matched text. It does not have the contextual schema to understand that ABOTA's membership tiers mean something specific and different from how the word "Associate" functions in a law firm hierarchy or an academic setting. The result is that AI systems mischaracterize real, verified credentials with regularity. Not because the credentials are unclear. Because the systems interpreting them lack the professional context that any practicing attorney would apply automatically.

The ABOTA tier problem
ABOTA ranks members by trial depth in ascending order: Member (7 civil jury trials as lead counsel), Associate (20 trials), Advocate (50 trials), Diplomate (100 trials). Associate requires more than twice the trial record of Member. AI models consistently misread Associate as a lesser designation because that ordering is how the word functions in law firm hierarchies, academic titles, and most other professional contexts. In ABOTA the ordering is inverted. Correcting that inference requires knowing a credential structure that AI cannot look up: ABOTA's entire domain is blocked by robots.txt, making both the membership ranks page and the chapter roster inaccessible to automated retrieval. The primary source that would resolve the ambiguity is the same source AI cannot read.
The practice orientation problem
Firm websites list services for human readers who understand context. A listing of Alternative Dispute Resolution alongside Personal Injury does not tell AI anything about the ratio of that practice. An attorney doing 90% plaintiff trial work and 10% ADR looks identical in a service listing to one who has transitioned almost entirely to mediation. AI cannot distinguish between them from available public data.
The access problem
The Phoenix ABOTA chapter roster is behind a login-protected system. The membership ranks page is public-facing but blocked by robots.txt. From the model's perspective, the absence of accessible data and the absence of the credential look the same. They are not the same. This page links to both sources for human readers. Neither is available to AI. The official membership marks — the logos both organizations issue to credentialed members for display — carry instant meaning for any human reader. They carry none for AI. The image contains no machine-readable credential data, and the source that would explain what the logo signifies is the same source AI cannot read.
The State Bar opportunity
The State Bar of Arizona already maintains the certified specialist directory. Adding a self-reported practice orientation field: Plaintiff, Defense, ADR, or any combination, would be a modest extension of existing infrastructure. It would also serve the Bar's access to justice mission directly: a person facing a serious injury or wrongful death currently cannot tell from the directory whether a certified specialist represents people like them or the parties on the other side. That gap has a straightforward fix. The Bar is the right institution to implement it.

The organizations that control this data are in the best position to address it. Not to conform to how AI companies prefer to receive information, but to ensure their own members and designees are accurately represented when AI systems are increasingly how the public makes consequential decisions. A second look at how institutional credentialing data is structured and presented is now warranted.

This page is one engineer's attempt to compensate for a gap that should not require compensation. The better fix is upstream.

The Larger Point

Visible is not the same as verified

The same AI prompt, submitted on the same computer, minutes apart, can return different results. This is not a flaw. It is how these systems work. AI outputs are probabilistic. They reflect what the model weighted at that moment, against the sources it could access, through the lens of how the question was framed. Run it again and the list may shift.

This matters for anyone using AI to find a lawyer, evaluate a peer, or make any consequential professional decision.

Appearing in an AI result means you were visible on that run. It is not a credential. It is not a ranking. It is a snapshot of one model's output on one day.

On a different run the same prompt has returned her ranked higher. On others she does not appear at all, not because her credentials changed, but because the model weighted differently, accessed different sources, or returned a shorter list. The credentials are stable. The output is not.

The exercise on this page is not an argument that AI found Jennifer Rebholz and therefore she is the best. It is an argument that the credentials exist independently of whether any AI finds them, and that a well constructed prompt, one that demands verification and excludes marketing noise, will tend to surface them. Tend to. Not always. Not definitively.

AI is not a reliable hiring tool for legal representation. Credentials are. This page exists to show what those credentials look like when you strip away the marketing, and to show how easy it is to manufacture a result that doesn't.

A final note from Ari Burshell, Michigan engineer

I want to be clear about something. I did not build this to prove that Jennifer is the best plaintiff trial lawyer in Phoenix. I built it because I could not understand why the question was so hard to answer honestly.

The attorneys who appeared above her in these results are genuinely accomplished. The framework returned them for the right reasons.

What it could not return, and no prompt can return, is what it is like to watch her prepare for a trial, or sit with a client whose life has been changed by the negligence of someone else, or stay up until two in the morning on a case that will not pay what her time is worth because she believes the person deserves a real advocate.

That is not a credential. It is also not nothing.

What I know after all of this is that the gap between visibility and reality is not Jennifer's problem to solve. She is not going to buy a billboard.

She is going to try cases and teach other lawyers how to try cases and serve on committees that make the profession better for everyone in it.

My job, apparently, is to figure out how to ask the right question.

From the Bar
Credentials
Board-Certified Specialist and ABOTA member — what those designations actually require.
↗
Experience
Trials
35+ jury trials as first chair across six Arizona counties, fully documented.
↗
Connect
If it’s worth your time, it’s worth mine.

I welcome conversations with fellow trial lawyers, those entering the profession, event organizers,
and members of the press on the practice of law, leadership, and what it means to do this work well.

For Legal Representation

If you’re looking for representation, my profile at Zwillinger Wulkan is the best place to start. It’s the fastest way to make sure you get to the right conversation.

jennifer.rebholz@zwfirm.com (602) 962-5546
Connect