The State of Legal AI: 2026 H1 Report.

Three shifts converged in ninety days.

The first half of 2026 is when legal AI stopped being a policy debate and started being a compliance reality. Three independent shifts — a wave of state bar action, an adoption rate that has outrun governance by more than two-to-one, and the arrival of private equity in personal injury — converged inside a single ninety-day window. Each is significant on its own. Together, they redefine what a serious law firm looks like in 2026.

This report is our attempt to make that picture legible. It is grounded in roughly thirty-five state bar opinions and proposals, twelve industry reports, public disclosures from the H1 2026 PE deals in personal injury law, and early observations from initial Counselcraft diagnostic engagements. We have tried to be specific where specificity is available, and honest where it isn't.

Three findings anchor everything that follows.

Each finding has its own section below. The forecast in Section 7 names six specific things we expect to see by the end of 2026. The recommendations in Section 8 are sized by firm headcount because the right answer for a ten-attorney firm is not the right answer for one with fifty.

If you read only one section after this one, read Section 4. The adoption gap is the throughline.

What we analyzed, what we didn't.

This is the first edition of what we intend to publish twice a year. As a first edition, it leans on synthesis of existing primary sources rather than original survey data. We want to be explicit about what's in scope and what isn't, because the credibility of the rest of the report depends on it.

Sources analyzed

Approximately thirty-five state bar formal opinions, advisory opinions, task force reports, and proposed rule amendments published between November 2023 and May 2026 — the full set of US jurisdictions that have issued substantive guidance on generative AI in legal practice, plus the ABA Formal Opinion 512 (July 2024) baseline. Twelve industry research reports from Clio (the 2025 Legal Trends Report, the 2026 Legal Trends for Mid-Sized Law Firms, the 2026 UK & Ireland Legal Insights Report), Thomson Reuters, the ABA AI Task Force, Gartner, McKinsey-BrightEdge AI search data, and recent Statcounter and Goodie analyses. Public disclosures and trade press coverage of three H1 2026 PE deals in personal injury law. Public coverage of Verisk's January 2026 standardized AI exclusion forms and the bar-affiliated malpractice carrier guidance from nine carriers (ALPS, OBLIC, TLIE, ISBA Mutual, Lawyers Mutual NC, FLMIC, Oregon PLF, LMICK, ALAS).

Primary observations

We have also incorporated early observations from initial Counselcraft diagnostic engagements. These are pattern-level findings, not statistical claims — we name them as "patterns we see" rather than "X% of firms" wherever the empirical base is too small to support a percentage. We expect the H2 2026 edition to add a planned 250-firm primary survey, which will allow us to convert pattern statements into adoption percentages with confidence intervals.

What this report is not

It is not a legal opinion. We are not advising on any specific bar rule, malpractice policy, or ethics question. Every firm should consult its retained ethics counsel for jurisdiction-specific obligations and its broker for coverage questions. We are operators describing what we see, not lawyers issuing guidance.

From silence to enforceable code in thirty months.

The regulatory picture in 2023 was simple: there wasn't one. Mata v. Avianca had just made fake AI-generated citations a national news story. The ABA Task Force on Law and Artificial Intelligence had just been formed. California became the first state to publish practical guidance — on November 16, 2023, in the form of a "living document" with no disciplinary weight. Beyond that, jurisdictions were silent.

The picture in mid-2026 is the opposite. Roughly thirty jurisdictions have issued substantive guidance. Two states have moved from advisory guidance toward enforceable rule changes. Multiple federal courts have imposed AI disclosure requirements on filings. Insurance carriers have begun pricing AI risk into renewal questionnaires. The trajectory is not slowing.

We'll walk through the four most consequential developments — the ones that change what a defensible posture actually requires — and then map the broader landscape.

ABA Formal Opinion 512 — the baseline (July 2024)

The ABA Standing Committee on Ethics and Professional Responsibility issued Formal Opinion 512 in late July 2024. The opinion does not create new rules. It interprets existing Model Rules in the context of generative AI: Rule 1.1 (competence) requires lawyers to understand the AI tools they use, including limitations and risk of hallucinations. Rule 1.6 (confidentiality) requires that client information not be shared with AI tools whose terms permit training on that data. Rule 5.1 and 5.3 (supervision) require that lawyers supervise both subordinate lawyers and non-lawyer assistance, including AI agents. Rule 1.5 (fees) requires that AI-driven efficiency be reflected in reasonable fees rather than billed as if the time had been spent.

Opinion 512 is the closest thing the United States has to a national baseline. Every state opinion that has followed it cites it, builds on it, or operates parallel to it. A firm following Opinion 512 will satisfy the floor in most jurisdictions. A firm not following it is increasingly indefensible in any jurisdiction.

California Rule 1.1 amendments — the inflection point (March 13, 2026)

The single most important regulatory event of H1 2026 is the California Committee on Professional Responsibility and Conduct's approval of proposed amendments to six existing Rules of Professional Conduct, opened for a 45-day comment period on March 13, 2026. If finalized, these amendments would make California the first US jurisdiction to embed AI-specific obligations into the enforceable disciplinary code rather than into advisory guidance.

The most significant amendment is a new comment to Rule 1.1 (Competence) explicitly addressing AI-generated citations. The proposed language clarifies that AI-generated citations are not exempt from the lawyer's verification duty, and that the verification obligation extends specifically to fabricated, misstated, or decontextualized authority. Five other rules — including Rule 1.6 on confidentiality and Rule 5.3 on supervision of non-lawyer assistance — receive parallel AI-specific language.

The shift matters because of what it changes about enforcement. California's existing 2023 practical guidance, despite being detailed and frequently cited, has no binding authority. The proposed amendments would make AI-specific failures sanctionable under specific, explicit authority. The cases will be easier to bring, the standards easier to enforce, and the defenses harder to mount.

We expect California to finalize these amendments by the end of 2026. We further expect that other large jurisdictions — likely New York, Florida, and Texas — will follow with rule amendments rather than continued reliance on advisory opinions. By the end of 2027, the bar of practice in the largest US legal markets will be set by enforceable rules, not by guidance documents.

Oregon Formal Opinion 2026-208 — the intake question (February 2026)

The Oregon State Bar issued Formal Opinion 2026-208 in February 2026, addressing whether law firms may use autonomous AI agents to handle client intake. The opinion's headline answer is "yes, qualified" — but the analytic structure of the opinion is what makes it the leading example of a new generation of bar action.

The opinion frames the use of autonomous AI intake under Rule of Professional Conduct 5.3, which governs lawyer supervision of non-lawyer assistance. The crucial move: the lawyer's duty is treated as supervision of the service provider — the vendor — rather than supervision of the AI agent itself. The AI agent is not an "assistant" under the rule's plain language; instead, the firm is responsible for vetting the vendor, understanding the data flow, ensuring confidentiality, and confirming the agent's behavior aligns with the lawyer's professional obligations.

The practical implications are substantial. A law firm using an autonomous AI intake tool in Oregon now needs documented vendor due diligence (terms, data retention, training-on-customer-data clauses, breach notification), a documented theory of how the AI's behavior is supervised, a documented disclosure to prospective clients that they may be interacting with AI, and a conflict-screening process designed into the system rather than bolted on after the fact.

In re Mostafavi — the discipline (February 2026)

For firms still treating AI bar discipline as theoretical: in February 2026, a California State Bar Court judge approved a disciplinary agreement against attorney Amir Mostafavi, whose briefs contained AI-generated fake quotations. The agreement includes a stayed one-year license suspension, required ethics school attendance, and — for the first time in a published US discipline matter — mandatory coursework specifically on the risks and benefits of AI tools in legal work.

Mostafavi matters because of what the sanction includes. AI-specific CLE as a remediation requirement signals that bar courts now consider AI-specific competence a distinct skill, separate from general legal competence. The implication is that "I didn't know" is no longer a defense.

The broader landscape

Beyond the four anchor events, the H1 2026 regulatory landscape spans every major jurisdiction. The matrix below maps it. Filter by status to see where each state sits.

The trajectory

The pattern across H1 2026 is clear. State bar action is moving along three dimensions simultaneously: from advisory to enforceable, from general to topic-specific (intake, citations, billing), and from reactive to anticipatory. The firms that build governance now will be unremarkable in eighteen months. The firms that wait will be the test cases.

Adoption is at 79%. Governance is at 47%.

The single most consequential statistic in legal AI in 2026 is the gap between adoption and governance. Both numbers come from the same source — Clio's Legal Trends Report — and both are confirmed by parallel research from the ABA, Thomson Reuters, and our own diagnostic work. The gap is enormous, it is widening, and it is the throughline of every other risk in the report.

The numbers

Clio's 2025 Legal Trends Report, drawn from a survey of thousands of US legal professionals, documents AI adoption rising from 19% in 2023 to 79% in 2025 — a more than four-fold increase in two years. The 2026 Legal Trends for Mid-Sized Law Firms report puts adoption at 86% in mid-sized firms. Among firms with 200 or more lawyers, adoption is at 87%. Even solo practitioners, the slowest-adopting segment, are at 71%.

In the same population, the governance numbers are inverted. 53% of legal professionals report that their firm has no AI policy or that they are unaware of one. Of the 47% that do have a policy, only 30% report a policy that allows and encourages AI use, 12% allow but do not encourage, and 5% prohibit. 44% of firms have no formal AI governance framework of any kind — no vendor whitelist, no documented review process, no named owner.

What this looks like in a real firm

We see the gap in every diagnostic we run. The specific pattern: a managing partner believes the firm uses "a couple of AI tools" — usually meaning a research tool and maybe a contract review platform that the firm formally subscribes to. The diagnostic uncovers three to five additional tools in active use that the partner is not aware of. A paralegal is using a free consumer LLM to summarize voicemails. An associate is pasting client communications into ChatGPT to draft initial response language. An intake coordinator is using a generic AI chatbot to handle off-hours inquiries. The marketing team is using an AI writing tool with terms that explicitly permit training on submitted content.

None of these uses is unreasonable on its own. Each represents a staff member trying to be efficient. The aggregate is a firm with five AI tools in active use, no policy, no whitelist, no audit trail, and no documented theory of how client confidentiality is preserved across any of them. When we ask the partner where client data goes for each of those five tools, the partner cannot answer for any of them. This is not a hypothetical. It is the consistent finding.

The vendor reality

Forty percent of legal professionals are using legal-specific AI solutions, according to Clio. The other 60% are using either general-purpose enterprise AI or — far more often — consumer-grade AI tools that were not designed for legal use cases. The distinction matters because the terms of service, data retention practices, training-on-customer-data clauses, and breach notification commitments are materially different between consumer and enterprise tiers of the same product.

Most firms are operating under the consumer terms of products they assume are operating under enterprise terms. A free or low-cost ChatGPT or Claude or Gemini account permits the vendor, under most current terms, to use submitted content to train future models. A paid enterprise account does not. A firm whose staff are submitting client information into the consumer tier is, in most jurisdictions, in violation of Rule 1.6 confidentiality obligations regardless of whether the bar has issued AI-specific guidance.

The insurance pressure

Malpractice carriers have moved faster than most state bars. According to ABA data, more than 60% of carriers now ask about AI use on intake or renewal applications. Specialty broker Jencap reported in IA Magazine (March 2026) that underwriters are routinely asking "How are you using AI?" on renewals, and that firms without documented protocols are projected to lose preferred-risk standing within eighteen months.

The largest standardized policy form provider in the US insurance market, Verisk, began rolling out new general liability AI exclusions in January 2026. Because insurers nationwide adopt Verisk templates as a default, this single action is reshaping market-wide coverage. Major commercial carriers — AIG, Berkley, Hamilton — are independently rolling out their own AI exclusions or sub-limits. A policy with a $10M face amount may cap AI-related losses at $500K, creating significant uninsured exposure for firms relying heavily on AI tools.

Nine bar-affiliated mutual insurance carriers — ALPS, OBLIC, TLIE, ISBA Mutual, Lawyers Mutual NC, FLMIC, the Oregon PLF, LMICK, and ALAS — have all published lawyer-specific AI risk guidance between 2023 and 2026. ALAS, the country's largest lawyer-owned mutual covering major firms, issued a bulletin titled "ChatGPT: Not Ready for Prime Time" explicitly warning policyholders that generative AI use could result in malpractice claims and that coverage under existing professional liability policies may not apply.

This is the most significant external pressure on firm AI governance today. Bar opinions can be ignored until a discipline case arrives. An insurance renewal cannot be ignored. The carrier asks for documentation; the firm either has it or it doesn't.

Why the gap is widening

Three structural reasons the adoption/governance gap is getting worse, not better:

One — adoption is decentralized; governance has to be centralized. Every individual staff member can adopt a new AI tool in five minutes. Governance requires a person, a process, and a written artifact. Adoption scales horizontally and frictionlessly. Governance scales vertically and with effort. The two run on different clocks.

Two — vendors are moving faster than firms. The major AI labs ship significant model upgrades on roughly quarterly cycles. Each upgrade meaningfully changes the capability and risk profile of the tools built on top of them. Without a quarterly governance cadence, the diligence goes stale faster than firms can refresh it.

Three — the legal AI market is fragmenting. In 2024, "law firm AI strategy" plausibly meant choosing a single research platform. In 2026, it means choosing among hundreds of category-specific tools across research, drafting, intake, contract review, voicemail transcription, document summarization, citation verification, client communications, billing automation, and conflict checks. Each category has multiple credible vendors. Each firm now has to make governance decisions across what is functionally a vendor portfolio.

The implication

The adoption/governance gap is the unifying mechanism behind almost every other risk in this report. It produces the failure modes in Section 5. It is the gating issue for the patterns in Section 6. It is the principal target of the regulatory action in Section 3. It is the lever PE is pulling on in Section 7. If a firm makes one operational change in 2026, this is the one to make.

Five failure modes, mapped by severity and prevalence.

Section 4 described the gap in aggregate. Section 5 names the specific failure modes that gap produces. These are the patterns we see surface in diagnostic engagements, ranked roughly in order of how often they appear and how much revenue or risk they touch.

Failure mode 1 — Tool sprawl without governance

Three to five AI tools in active use across the firm, no formal whitelist, no documented vendor terms review, no clear owner. Adoption happened bottom-up, one staff member at a time. No one at the partner level can produce a current inventory.

The operational impact: every tool in use is a potential confidentiality breach, a potential malpractice exclusion, and a potential bar discipline trigger. The firm cannot pass a carrier renewal questionnaire honestly. The firm cannot respond to a bar inquiry credibly. The firm cannot answer a client question about AI use truthfully because no one knows the answer.

Failure mode 2 — Intake AI without supervision

A chatbot, voicemail summarizer, or autonomous intake agent is handling some portion of prospective client contact, with no documented theory of how the lawyer is supervising the tool, no disclosure to prospective clients that they may be interacting with AI, and no conflict-screening designed into the system.

Post-Oregon Formal Opinion 2026-208, this is the failure mode most likely to produce a bar discipline case in 2026. Twenty-four-seven AI intake is, in many ways, the most operationally valuable AI deployment a firm can make. But the same deployment without supervision is the highest-exposure use case in the firm.

Failure mode 3 — Citation verification gaps

Associates and paralegals use AI tools to draft work product. The drafts include cited authority. The supervising attorney signs off on the work product without independently verifying that every cited case, statute, and rule exists, is correctly stated, and supports the proposition for which it is cited.

This is the Mata v. Avianca failure mode and the Mostafavi failure mode. California's proposed Rule 1.1 amendments make the verification obligation explicit. Most other jurisdictions reach the same result under existing competence rules.

Failure mode 4 — Marketing channel optimization for the wrong outcome

Paid media campaigns optimize against form submissions, calls placed, or other shallow conversion events. The firm celebrates falling cost-per-conversion numbers while the actual cost-per-signed-case rises. The failure mode is industry-wide and has worsened in 2026 as Performance Max and similar platform-driven optimization tools have become defaults.

Failure mode 5 — The partner bottleneck

The managing partner is the operating system of the firm. Every meaningful decision routes through her. The firm has no documented SOPs, no KPI dashboard, no decision rights structure beneath partner level, no documented intake-to-retainer process.

This failure mode is upstream of every other one. A firm with a partner bottleneck cannot install governance, cannot supervise AI intake, cannot build citation verification workflows, and cannot run marketing attribution — because the partner has no time and no one else has authority.

Four patterns that reinforce each other.

The mirror of Section 5. Four patterns we observe in firms that have moved out of the failure modes. They are not the four hardest problems to solve. They are the four problems whose solutions reinforce each other most.

Pattern 1 — Four-layer marketing attribution

Firms that are growing efficiently in 2026 share one architectural feature: every signed retainer ties back to a tracked source through a documented chain. The four layers are tracked entry points, identity stitching across sessions, stage transitions from inquiry to retainer, and financial reconciliation against billing. Most firms have layer one. Few have all four. The firms that do can answer the question "which marketing dollar produced this signed case" in dollars and days. PE buyers want this. Carriers don't ask for it but increasingly use it as a proxy for operational maturity.

Pattern 2 — Oregon-compliant intake stack

Firms running AI-powered intake successfully share the same configuration. The AI agent handles initial capture and triage. A documented vendor due diligence file covers data flow, retention, and supervision theory. A clear disclosure tells prospective clients they may be interacting with AI. Conflict screening is designed into the system, not bolted on. A lawyer reviews and approves all engagement-creating communications before any attorney-client relationship is established.

This configuration satisfies Oregon Formal Opinion 2026-208 explicitly and ABA Formal Opinion 512 implicitly. It also happens to be the configuration that produces the best operational outcomes — fewer drops at the 24-hour mark, higher conversion to consult, cleaner downstream matter quality. The compliance posture and the growth posture converge.

Pattern 3 — Quarterly governance cadence

The firms that maintain coherent AI governance over time have a recurring meeting on the calendar. Sixty to ninety minutes, every quarter. The COO or designated governance owner runs it. Standing agenda: review the tool inventory, review new tools requested by staff, review the policy against any new bar opinions or rule changes, document any updates.

This is the lowest-effort, highest-leverage AI governance practice we observe. The cost is four hours per year of the governance owner's time. The benefit is the ability to honestly answer "yes" when a carrier asks about ongoing AI governance.

Pattern 4 — PE-readiness as a maturity signal

The most interesting pattern we see is not in firms that are selling to PE — it's in firms that are not. Firms that have done the work to be acquirable have, by definition, built the systems Section 5 says are missing in the failure-mode firms.

We see firms with no intention of selling adopting this posture because the same systems that make a firm acquirable make it competitively durable. A PE-backed competitor in your market can afford to lose money on marketing for eighteen months to take share. They can pay above-market for the best paralegal in your office. They can roll out a 24/7 AI intake stack before you can. A firm whose own operations match the documentation level of a PE-backed competitor can defend its market position.

The Clio 2026 data offers a corroborating signal. Mid-sized firms (25–50 attorneys) are pulling ahead of both ends of the market — growing revenue 4× faster than headcount, expanding the work they can take on, reporting the largest gains in client satisfaction.

Six predictions for H2 2026.

We will revisit each in the H2 2026 edition of this report. The predictions are specific enough to be graded.

The work, sized by firm.

The right answer depends on the firm's size, complexity, and current maturity. Three buckets follow. Pick the one that fits your firm and start at the top of the list.

1–10 attorneys: minimum viable governance

The goal is to be defensible, not exhaustive. Three weeks of focused work produces the artifacts that satisfy carrier renewal questions, bar inquiries, and client transparency expectations.

1. Write a one-page AI policy. Approved tools, prohibited behaviors, named policy owner. Dated and signed.
2. Inventory every AI tool in use — including the ones the partner does not currently know about. Ask staff directly.
3. Establish vendor terms review. Use the enterprise tier of any consumer-facing product. Do not allow client information into free or low-cost consumer tiers.
4. Build a citation verification workflow. Every piece of AI-touched work product gets a documented attestation.
5. Update malpractice renewal disclosures. Pre-empt the carrier question by having the documentation ready.

Time: ~2-3 weeks of focused work, one named owner. Cost: minimal.

10–50 attorneys: the seven priorities

This is the segment where the failure modes compound fastest and where the working patterns produce the largest economic gains.

1. Execute all five 1–10 items first. Foundation before structure.
2. Build four-layer marketing attribution. Unlocks honest cost-per-signed-case measurement.
3. Deploy Oregon-compliant intake. Solves the 24-hour intake gap and the bar compliance question simultaneously.
4. Establish quarterly governance cadence. Recurring meeting, named owner, four hours per year of leadership time.
5. Document SOPs for intake, conflict checks, matter opening, and billing. Functionally, to the level where a new hire could execute from the document.
6. Unhook the partner from the operating system. Decision rights below partner level, KPI dashboard, fractional or full-time COO if warranted.
7. Audit AI search visibility quarterly across ChatGPT, Gemini, Perplexity, and Claude.

Time: ~12-16 weeks of focused work, distributed across team.

50+ attorneys: the PE-readiness checklist

Whether you intend to sell, intend to acquire, or intend to remain independent, the work is the same. Six artifacts a buyer's analyst would ask for — that a defensively positioned firm should have regardless:

1. A 36-month cohort analysis of marketing spend by channel against signed cases and case value. Weekly granularity for the most recent 12 months.
2. An intake yield report by channel and by hour-of-day. 12 months of history.
3. A documented SOP library. Versioned, with named owners.
4. A KPI dashboard reviewed weekly at leadership level. Named owners for each metric.
5. A written AI policy with vendor whitelist, quarterly governance review, and training records. Non-optional.
6. A 12-month forward marketing plan with budgets, channel mix, projected signed-case yields by month, and a documented attribution methodology.

A firm with all six is in the top 5% of US legal practices on operational maturity. A firm with none of them is the firm PE buyers either discount heavily or walk away from — and equivalently, the firm that loses ground to PE-backed competitors over the next thirty-six months.

Sources, authors, and the next edition.

How to engage Counselcraft

Every Counselcraft engagement begins with a paid Growth & Operations Diagnostic. The diagnostic runs two to three weeks, includes both founders personally, and produces a written assessment of the firm's current state across the five domains in this report: demand generation, intake, operations, attribution, and AI readiness. The deliverable is yours regardless of whether the engagement continues. If you decide to move forward into implementation, the diagnostic fee credits toward the implementation engagement.

To inquire: counselcraft.legal/contact or hello@counselcraft.legal.

About the authors

Bob Clary is the co-founder of Counselcraft, leading the growth side of the business. He has spent twenty years building demand engines for law firms — paid media, SEO, content, attribution, and AI-search optimization. His prior agency made the Inc. 5000 list fourteen times. He is based in Syracuse, NY.

Jim Firenze is the co-founder of Counselcraft, leading operations. He has spent decades running operations at scale across customer service, technical services, and compliance, including as a senior operations leader at a $100M business that exited to private equity for over $300M. He is the sitting Chief Operating Officer at a leading criminal defense firm and is based in Syracuse, NY.

Citations

Section 3 — The Regulatory Landscape: ABA Formal Opinion 512 (July 2024); California State Bar COPRAC Proposed Rule Amendments (March 13, 2026); California State Bar Practical Guidance (November 16, 2023; revised May 14, 2026); Oregon State Bar Formal Opinion 2026-208 (February 2026); Oregon State Bar Formal Opinion 2025-205; In re Mostafavi, California State Bar Court (February 2026); Mata v. Avianca, S.D.N.Y. (2023); State Bar of Texas Ethics Opinion 705 (February 2025); Texas AI Task Force Preliminary Report (February 2024); Florida Bar Ethics Opinion 24-1; NYSBA Task Force on AI Report (April 2024); NYC Bar Formal Opinions 2024-5 and 2025-X; 22 NYCRR Part 161; North Carolina 2024 Formal Ethics Opinion 1; Pennsylvania Joint Formal Opinion 2024-200; Washington State Bar Advisory Opinion 202505; Connecticut Public Act 25-113.

Section 4 — The Adoption Gap: Clio 2025 Legal Trends Report; Clio 2026 Legal Trends for Mid-Sized Law Firms; Clio 2026 UK & Ireland Legal Insights Report; ABA AI Task Force Year Two Report (2025); Thomson Reuters Future of Professionals Report; Verisk January 2026 standardized AI exclusion forms; ALAS "ChatGPT: Not Ready for Prime Time" policyholder bulletin; IA Magazine Jencap underwriter survey (March 2026); carrier-published AI risk guidance from ALPS, OBLIC, TLIE, ISBA Mutual, Lawyers Mutual NC, FLMIC, Oregon PLF, LMICK, ALAS.

Sections 5–7 — Failure Modes, Patterns, Forecast: Counselcraft diagnostic engagements (aggregated patterns, identifying details removed); Statcounter / Aodhan Cullen AI referral traffic market share (March 2026); Goodie 2026 AI Search Traffic Report; Adobe AI referral visit data (June 2025); Gartner agentic AI initiative failure rate projection (2025); McKinsey / BrightEdge AI search query routing analysis; Bloomberg Law coverage of Dudley DeBosier / Uplift Investors / Orion Legal (January 2026); Legal Futures coverage of Rafi Law Group / Rafi Law Services MSO (April 2026); Bloomberg Law coverage of Hughes & Coleman / Uplift Investors / Orion Legal (May 2026); Private Equity Legal Alliance white paper (January 2026); CLS Blue Sky Blog (April 30, 2026).

Disclosures

This report is informational and does not constitute legal advice or insurance advice. Firms evaluating their own bar compliance posture should consult retained ethics counsel for jurisdiction-specific obligations. Firms evaluating their own malpractice coverage should consult their broker, retained coverage counsel, or both. The authors do not have a financial interest in any of the AI vendors, malpractice carriers, or law firms named in this report. Counselcraft is not affiliated with any of the bar associations, courts, or regulatory bodies referenced.

Edition note

This is the H1 2026 edition. The next edition will be published in approximately six months and will include a planned 250-firm primary survey to convert pattern observations into adoption percentages with confidence intervals.