Insight · General counsel hiring

Tools versus marketing: a new selection criterion.

The story is that every panel firm already has the tools. The number a New York general counsel can actually use is how many of those firms can be scored on tools versus marketing — and whether that score is a reason to hire from the firm or keep it on the panel.

Hire in-house counsel Submit your CV
01 Start here

Pick the thing you can actually score.

Sartori maps about 67,000 lawyers in New York. Coverage is not the constraint. A general counsel can name three vendors on a panel website and still fail to name who runs them on the company's matters. The advertised ai stack is not a hiring criterion until that name exists.

State 01 · Marketing page A vendor logo is not a stack you can score.

The firm has a public AI page. No named operator, no written confidentiality architecture, no matter-level yes or no. A chief legal officer who hires from that page is buying marketing. Do not treat this as a criterion.

The contradiction is not that firms lack tools. It is that a general counsel cannot tell which state a given firm is in. What you can actually score is laid out below.

36%
CLOs already in active GenAI deployment1,049 chief legal officers, 43 countries
ACC Chief Legal Officers Survey, January 2026
<1/3
corporate legal and tax departments that know if firms use AI on their work1,500+ professionals, 27 countries
Thomson Reuters Institute, 9 February 2026
85%
departments with a dedicated AI resource or committee135 departments, median revenue $13 billion
Harbor / CLOC, December 2025
48%
firms with a vetted-only generative-AI policyILTA 2025 sample of 580 firms
Law360 Pulse reporting ILTA, 2025
02 The gap

How many firms a general counsel can actually score.

The marketed story is that every large firm already has the tools. The buyer's number is how many of those tools can be named, governed and checked on the company's own matters.

We have worked the New York in-house market for more than 10 years, for general counsel, chief legal officers and heads of legal in financial services, healthcare, insurance and technology. A marketed vendor is not a stack a general counsel can score.

Across the same New York interview cohort, 241 of 412 general counsel, chief legal officers, deputies and heads of legal operations over a 24-month window said they could not name whether their three largest panel firms used a sanctioned generative tool on the company’s matters. That cut sits inside the one New York interview book. It is the in-house read of a public fog the Association of Corporate Counsel and Everlaw measured in October 2025: most legal departments are not requiring the tools, and most cannot tell whether the firms are using them.

Eighty percent of departments in that 2025 ACC/Everlaw survey of 657 in-house professionals across 30 countries neither require nor encourage outside counsel to use generative AI. Adoption is driven by the in-house team itself in 58 percent of responses; law firms leading the charge: 1 percent; truly collaborative: 3 percent. Among companies with revenue of $10 billion or more, outside-counsel guidelines set AI expectations in 39 percent of cases. Thomson Reuters Institute, in the 9 February 2026 article on its 2026 AI in Professional Services Report, added the other half of the fog: 40 percent of firm respondents have received orders both to use AI and not to use AI from various clients. The landing page for that report puts two-thirds of corporate respondents wanting firms to use AI, and fewer than 20 percent mandating it.

ILTA’s 580-firm sample is the universe of technology leaders a New York general counsel is trying to score, not a census of every panel. Tool presence is not deployment. Law360 Pulse’s 2025 read of that survey is the mix in the chart below. Those shares say a product is in the building. They do not say a lawyer on your matter has it, has been trained on it, or is forbidden from pasting your facts into a public chatbot.

The American Bar Association Law Practice Division’s Legal Industry Report 2025, a survey of over 2,800 legal professionals, puts a thinner stack under that ILTA headline: personal generative AI use at work at 31 percent, firm-wide generative AI at 21 percent. Thomson Reuters Institute’s 2026 professional-services sample puts organization-wide generative AI at 40 percent, up from 22 percent the prior year. Different instruments, different years, same direction: individual use runs ahead of a firm-wide stack a chief legal officer can audit. A logo is not a stack a general counsel can score.

A chief legal officer at a New York-headquartered financial-services group told us, in that same 24-month interview window, that the panel websites named three vendors and none of the humans who ran them. The Association of Corporate Counsel publishes a first-pass instrument for that conversation — Top 10 GenAI Transparency & Readiness Questions for Outside Counsel — covering data security, professional responsibility, quality control and cost transparency. The question list itself sits behind a member login. The purpose does not: named accountability, opt-in controls, systematic quality checks, transparent billing. Until those answers exist, the firm is in decoder state one.

Share of ILTA 2025 respondent firms reporting each product, not the share of lawyers who have it on a given matter. Presence is not deployment.

Law360 Pulse reporting the ILTA 2025 Technology Survey (580 firms).

A logo is not a stack a general counsel can score.
On what you can score
03 The valve

The company is pulling work in. Headcount is not the release.

A general counsel who treats a firm's tools as a hiring criterion is making that decision inside a budget that is trying to insource drafting and to freeze the lawyer count.

The Association of Corporate Counsel and Everlaw, in the 14 October 2025 release, found 64 percent of in-house respondents expect to rely less on outside counsel, and 50 percent expect reduced outside-counsel costs. Insourcing intent is not even across the docket; the mix is in the chart below. Harbor and CLOC, across 135 departments with median revenue of $13 billion, found 65 percent had intentionally kept work in-house in the last one to two years. The work a firm’s tools are supposed to cheapen is the work the company is already trying to absorb.

That absorption is not a hiring wave. The 2026 ACC Chief Legal Officers Survey of 1,049 CLOs across 43 countries found 63 percent expect legal-department headcount to remain stable. Harbor put the share of departments expecting lawyer-headcount increases at 32 percent, down from 42 percent the prior year. When the hire happens, it is operations, vendor-management or a specialist AGC — not a bench of juniors to run prompts. Forty-seven percent of those CLOs said technology and AI proficiency is the primary skill their CEO wants them to develop. The stack-scoring job is being pulled onto the chief legal officer’s own description even where no vice president of legal operations exists.

Thomson Reuters’ Future of Professionals Report 2026, fielded March–April 2026 across 1,816 professionals in 62 countries, found 32 percent of corporate clients will reconsider provider relationships within 12 months. Among those reconsidering, one-third put more than $1 million in annual work at risk. Of the 37 percent of ACC/Everlaw respondents who said some or all of their firms use generative AI on their matters, 59 percent had seen no clear savings. Thomson Reuters Institute’s February 2026 professional-services report found only 18 percent of organizations track AI ROI, and 40 percent do not know if ROI is measured at all. A general counsel cannot score a saving the firm does not measure.

Our New York mandate telemetry on the in-house line records the same bind in the requisition. Of 24 closed searches over three years, 11 named tool-governance, vendor-management or AI-policy ownership as a duty on a general counsel, deputy, legal-operations or specialist AGC brief. Seven of those 11 stalled at least one cycle because the company could not describe the stack the hire would own. Four of those 11 still opened as “AI-fluent lawyer” with no named duty — a miss in our own intake, and the files that took the long end of the four-to-seven-month window. We cannot see a firm’s actual deployment from candidate interviews alone. That is a limit of the research, not a slogan about method.

A head of legal operations at a New York healthcare platform told us the last panel RFP returned a marketing one-pager when the company asked for an AI policy. Harvey’s 25 March 2026 announcement counted over 500 in-house legal teams among its customers — the buyer class for the company’s own stack, not a headcount census. ABA Journal, on 11 August 2026, reported a Law.com Q2 survey in which AI and innovation investments increased more than 5 percent at nearly 70 percent of Am Law 100 respondents. Capital is running ahead of the scored stack. The legal-tech CV question sits on a sibling page; this page is the buyer’s scoring problem.

Share of ACC/Everlaw in-house respondents who expect to insource each task with generative AI. High-judgment work is not where the expectation sits.

ACC/Everlaw generative-AI survey, released 14 October 2025 (n=657).

04 The seats

The scorer sits in the company. The operator sits in the firm.

A general counsel scoring a firm's tools is doing Firm and Vendor Management plus Technology, often through a legal-operations professional. Hiring from the firm is a different seat: the admitted lawyer who runs the tools inside a named practice.

Thomson Reuters’ 2025 Legal Department Operations Index, fielded in July among more than 125 corporate legal-department professionals, found 81 to 82 percent of departments have at least one dedicated legal-ops role — and 45 percent of respondents classified themselves as “General Counsel tasked to run legal operations.” Fifty-five percent reported flat or decreasing legal-department budgets; 51 percent reported flat legal-tech budgets. Among GC respondents, 47 percent were more focused on service enhancement than cost reduction, and 7 percent more focused on cost. CLOC’s Core 12 names the functions that own the score: Firm & Vendor Management, Technology, Knowledge Management, Service Delivery. The career path for that ops ladder is mapped separately. This page only needs the two functions that score a firm’s tools.

The Association of Corporate Counsel 2025 Law Department Compensation Survey, 1,632 U.S. self-reports effective 1 March 2025, is the title taxonomy a New York general counsel actually pays. The cash ladder is on the scale below; the figures are national self-reports, not a Manhattan premium. The 2026 ACC CLO survey puts the share using the title Chief Legal Officer at 34 percent. ABA Journal, reporting the 2025 National Lawyer Population Survey on 9 December 2025, put New York at 190,015 resident active lawyers. That is a bar census, not a count of people who can govern a sanctioned tool.

Firm-side, the 2026 New York employer career pages that actually name the tools are practice-embedded, not innovation-theatre. An Am Law 100 global firm posted a New York Knowledge & Innovation Counsel seat on 2 September 2026 at $193,000 to $301,000, J.D. required, at least six years’ practice, assigned to intellectual property, regulatory (including antitrust and data privacy), private client and tax, restructuring, energy, real estate or financial institutions, and charged with guiding lawyers across Harvey, Claude, ChatGPT, CoCounsel and Copilot. Another Am Law 100 firm posted a New York Director, Knowledge Management & Innovation on 27 August 2026 at $260,000 to $390,000, J.D. required. A litigation-heavy Am Law 100 firm posted a New York KM Legal AI Specialist on 23 April 2026 at $135,000 to $165,000, J.D. required. A Legal Knowledge Engineer spec dated 26 February 2026 sits beside those lawyer seats: bachelor’s preferred in law, computer science or information systems, J.D. not required, curating evaluation datasets for transactional workflows. A general counsel hiring from a firm is usually buying the Counsel. A general counsel scoring a panel is asking whether the firm has the Counsel and the engineer and a written policy.

Where the work sits when a general counsel treats the tools as a hiring or panel criterion. The company scorer, the firm-side operator, and the non-lawyer engineer are three different seats.
Seat Where it sits What the general counsel is buying Credential gate
General counsel / chief legal officer Company payroll The rule: require, permit or bar generative AI on company matters; own the ethics perimeter New York admission; board-facing
VP / director, legal operations Company payroll Firm & Vendor Management plus Technology: RFPs, panel reviews, vendor clauses Often the general counsel is this seat
Knowledge & Innovation Counsel Panel firm, practice-embedded The operator: curates know-how so approved tools can use it; trains the practice J.D.; six-plus years in the assigned practice
Legal Knowledge Engineer Panel firm, beside the lawyers Context libraries, evaluation datasets, quality benchmarks for transactional workflows Bachelor's; J.D. not required
U.S. in-house median base cash from the ACC 2025 Law Department Compensation Survey, data effective 1 March 2025. National self-reports, not a New York premium. Click a marker for the title.
ops VP to GC median base
$88k analyst$330k GC median base

Analyst, legal operations

Median annual base. The bottom of the ops ladder that often holds the score.

ACC 2025 Law Department Compensation Survey, effective 1 March 2025 ↗
05 The clocks

New York already made unsanctioned tools a legal problem.

The demand that a general counsel treat a firm's tools as a criterion is not a technology fad. It is a stack of dated New York and federal obligations that make an unaudited vendor a sanctions, examination and disclosure problem.

Mata v. Avianca, No. 22-cv-1461 (S.D.N.Y. 22 June 2023), is the federal template a New York general counsel uses when asking whether a firm’s stack includes a human gate. The court held that using a reliable artificial-intelligence tool is not inherently improper, and that Rule 11 still requires a gatekeeping review. Counsel had submitted non-existent opinions generated by a public chatbot. Sanctions: $5,000 jointly and severally into the court registry, plus letters to the client and to each judge falsely identified as author. Do not treat the individual respondents as a hiring story. Treat the holding: a professional-grade tool plus a human check, not a public chatbot.

Statewide, 22 NYCRR Part 161 — added 25 March 2026 by Administrative Order 75, effective 1 June 2026 — sets a Unified Court System policy the Legal AI Governance tracker quotes as: AI use in preparing court papers should not be prohibited if existing duties are met, and attorneys should not be required, on submission, to disclose AI use. Individual courts may adopt a model rule under which the signer who used AI must independently ensure the paper contains no fabricated authorities. Signature still certifies under 22 NYCRR 130-1.1. The Southern District commenced 11,109 civil cases in the twelve months ending 30 September 2025, on the Administrative Office of the U.S. Courts’ Table C. That is the docket on which Mata-style gatekeeping and Part 161-style verification bite. Ask who checks citations before filing, not whether the firm “has AI.”

On the company side the stack splits. NYDFS Insurance Circular Letter No. 7, dated 11 July 2024, pulls insurance-company general counsel into model governance, proxy assessments and vendor-contract terms for artificial-intelligence systems used in underwriting and pricing; the board is responsible for outcomes, and DFS may examine under Insurance Law sections 308 and 309. Reuters, on 15 November 2024, reported the Department’s 16 October 2024 industry letter on 23 NYCRR Part 500: no new requirements beyond Part 500, but examinations will ask whether risk assessments account for the entity’s AI and vendors’ AI, with multi-factor authentication and data-inventory dates already in 2025. New York City Local Law 144 of 2021, enforced by the Department of Consumer and Worker Protection from 5 July 2023, is the company’s HR stack — automated employment-decision tools, bias audits, notice to candidates. A general counsel who confuses that with a law firm’s legal tools will miss Local Law 144 while obsessing over a vendor logo.

The RAISE Act — original signature 19 December 2025, chapter amendment signed 27 March 2026, effective 1 January 2027, as Davis Polk summarized on 20 May 2026 and DWT on 14 April 2026 — puts DFS and the Attorney General over frontier-model developers trained above 1026 FLOPs. Large developers (prior-year gross revenue above $500 million) file with DFS on a 72-hour critical-incident clock. Companies that only license or host others’ models are not frontier developers. For the corporate reader that is a vendor-questionnaire item, not a reason to hire a Knowledge Counsel at a bank. FTC Operation AI Comply, announced 25 September 2024, and the SEC’s March 2024 AI-washing settlements ($400,000 combined, as Davis Polk restated on 3 October 2024) make overselling the tools themselves a legal risk. The European Union’s AI Act, Regulation (EU) 2024/1689, is the extra-territorial hook for New York general counsel of EU-facing groups; the deployer-versus-provider map belongs on the EU Act seat page, not here. Treat the advertised tool as marketing until you can name who runs it on your matters.

New York and federal clocks that turn a marketed tool into a scored duty. Sort by the live-since column to see which obligation is already examinable.
Instrument Live since What the general counsel asks Seat that moves
Mata v. Avianca (S.D.N.Y.) 22 June 2023 Who gatekeeps citations before a Southern District filing? Any lawyer who signs the paper; supervision on the firm
NYC Local Law 144 (AEDT) 5 July 2023 Is the company's hiring tool bias-audited? That is the HR stack, not the law-firm stack. Employment counsel / labor AGC
NYDFS Insurance Circular Letter No. 7 11 July 2024 Can we show an underwriting model is not unlawfully discriminatory, including vendor models? Insurance-regulatory counsel plus vendor-contract ops
ABA Formal Opinion 512 29 July 2024 Informed consent before client information enters a generative tool that is not adequately protected? General counsel; whoever owns outside-counsel guidelines
NYC Bar Formal Opinion 2024-5 August 2024 Does the legal department sit inside the same ethics perimeter as the panel firm? General counsel / chief legal officer
NYDFS Part 500 industry letter 16 October 2024 Do risk assessments and vendor contracts account for the company's AI and the vendor's AI? Bank / insurer / fintech GC plus cybersecurity counsel
22 NYCRR Part 161 1 June 2026 Who independently verifies New York state-court papers contain no fabricated authorities? In-house litigation / managing attorney; firm KM that productizes the check
RAISE Act (frontier developers) 1 January 2027 Is this model provider a frontier developer, or are we only a deployer? DFS-fluent counsel at covered labs; vendor questionnaire for everyone else
Treat the advertised tool as marketing until you can name who runs it on your matters.
On the criterion
06 The criterion

When an AI stack is a selection criterion.

The test is not whether a firm has a product. It is whether the company can name the operator, write the rule, and staff the seat that will govern both.

Of the 11 New York in-house files whose specifications named the tools, nine were general counsel, deputy or legal-operations seats and two were specialist AGC briefs (privacy, cyber, employment or insurance regulatory). Five of the 24 closed searches on this line were legal-operations or vendor-management seats; nine were general counsel, chief legal officer or deputy; four were specialist AGC; six were other in-house. Completion on the New York in-house line is 94 percent. Median offer-to-acceptance is 16 working days. Counter-offer incidence is 28 percent — stack-governance seats are not immune, because the incumbent’s current employer has often just bought a license too. When companies call legal recruiters New York desks on this brief, they are usually asking for a deputy general counsel or a legal-operations hire, not for a prompt engineer. Our in-house counsel search and legal-operations recruiting are the two desks that actually run these files.

Two composites from that book, described by category only. A New York-headquartered financial-services group opened a deputy general counsel search because it wanted to hire “from the firm that has Harvey.” Six panel firms had a vendor page; two could name a Knowledge Counsel the company could actually interview. We rewrote the brief around that operator test and closed in six months, inside the four-to-seven-month window. A regional insurer writing into Circular Letter No. 7 opened an “AI lawyer” requisition. The file that closed was insurance-regulatory counsel plus a legal-operations vendor-diligence seat. The prompt-engineer version of that brief would not have survived an examination. A third file, a healthcare platform hiring a legal-operations director to score vendors, stalled because the company had no written AI questions — one of the seven stalls inside the 11. We accepted that brief too early. The cheaper intervention is to write the questions before the search, not after the first shortlist.

Across London’s interview cohort of 750, in-house respondents described the same scoring-blindness. That is a contrast, not a second mandate book. New York City Bar’s August 2026 emerging-companies and venture-capital paper recommends designating a technology ethics officer and treating formation documents and NDAs as more suitable for assistance than financing papers and M&A. The general counsel scorecard is the seat underneath this criterion. The research program note is how the New York figures are coined.

For the panel, the stack is a review item, not a relationship lunch.

  • Write require, permit or bar into the outside-counsel guidelines before the RFP. Most departments still do neither; the large-company cohort that uses guidelines is the exception, not the base rate.
  • Put a named operator on the questionnaire. A product name without a human is decoder state one. ACC’s Top 10 instrument is the first-pass list; green flags are accountability, opt-in controls, quality checks, billing transparency.
  • Use the annual review you already run. Harbor/CLOC found 46 percent of departments run structured annual outside-counsel reviews, with another 18 percent planning them; 61 percent have completed or are implementing preferred-provider panels; 76 percent use alternative fees.
  • Ask who checks citations before a New York filing, and whether the fee reflects time the tool saved. Those are ethics questions, not innovation theatre.

For the payroll, the hire is ops or a specialist AGC, not a prompt engineer.

  • Name the duty in the specification. Tool-governance, vendor-management, DFS examination support, citation verification, Local Law 144 ownership. “AI-fluent” is how four of our 11 stack-named files slipped.
  • Hire the admitted operator if you are hiring from a firm — Knowledge Counsel, not the engineer — unless the brief is genuinely an evaluation-dataset seat.
  • Do not open the search until the company can describe the stack the hire will own. The clock that slips is the brief, not the offer letter.
  • Separate the HR tool from the legal tool. Local Law 144 is employment counsel. Circular Letter No. 7 is insurance-regulatory counsel. Part 161 is litigation. One “AI lawyer” requisition will not cover them.

Questions a New York general counsel asks about a firm's tools

Should a New York general counsel treat a firm's ai stack as a hiring criterion?

Only after the company can name who runs the tools on its matters: 59 percent of in-house respondents in the 2025 Association of Corporate Counsel and Everlaw survey did not know whether outside counsel used generative AI on their work. A vendor logo is marketing. A named operator plus a written confidentiality architecture is a criterion a chief legal officer can hire against or keep on the panel.

How many panel firms can a New York general counsel actually score on generative AI?

Fewer than the marketing pages imply: Thomson Reuters Institute, on 9 February 2026, found less than one-third of corporate legal and tax departments aware whether their firms used AI on client work. ILTA’s 2025 Technology Survey put 80 percent of 580 firms as using or exploring generative AI, with full deployment rare.

Who inside the company should own scoring a firm’s tools?

The general counsel, usually through legal operations: Thomson Reuters’ 2025 Legal Department Operations Index found 45 percent of respondents were a General Counsel tasked to run legal operations. Harbor’s 2025 Law Department Survey found 85 percent of 135 departments had a dedicated AI resource or committee — the seat that should hold Firm & Vendor Management plus Technology in CLOC’s Core 12.

What does the in-house seat that governs AI actually pay?

The Association of Corporate Counsel 2025 Law Department Compensation Survey, data effective 1 March 2025, puts the median General Counsel / Chief Legal Officer base at $330,000 and Vice President, Legal Operations at $246,000. Those are national self-reports, not a New York premium, and they describe the company payroll, not a firm-side Knowledge Counsel.

Is a Knowledge and Innovation Counsel the person a company should hire from a firm?

Usually, if the criterion is someone who can operate and govern the tools: J.D.-gated New York Am Law 100 career pages in 2026 ran from $135,000 to $390,000 across the Knowledge Counsel, KM Legal AI Specialist and KM director seats. The Legal Knowledge Engineer spec beside those seats did not require a J.D. A general counsel hiring onto the payroll is usually buying the admitted lawyer, not the engineer.

How long does a New York in-house search for this kind of seat take?

Budget four to seven months on Sartori’s New York in-house line. Median offer-to-acceptance is 16 working days once a candidate is in play; the clock that slips is the brief that still says “AI-fluent” without naming the duty the hire will own.

07 Sources

ACC and Everlaw surveys, ILTA deployment stages, New York ethics opinions, and the New York interview cohort.

Transparency and spend-valve figures come from ACC/Everlaw, Harbor/CLOC and Thomson Reuters. Deployment stages come from ILTA. The New York clocks come from the docket, DFS, the City Bar and Part 161. Sartori's quarterly survey has run since 2019 across a global mapping of nearly 1.5 million lawyer profiles; reachability on this page is the New York in-house cut of that program.

Sources and further reading

38 references
  1. Sartori & Partners — New York Legal Talent Research Programme (1,675 structured interviews; ~67,000 lawyers mapped; quarterly surveys since 2019; mandate telemetry) sartoriglobal.com ↗
  2. 64% of In-House Counsel Expect GenAI to Reduce Reliance on Outside Counsel, 50% Expect Lower Outside Counsel Costs, ACC/Everlaw Survey Finds everlaw.com ↗
  3. The Transparency Gap: In-house Counsel Lack Clarity on Law Firms’ AI Use corporatecounselnow.com ↗
  4. 2026 AI in Professional Services Report: AI adoption has hit critical mass, but now comes the tough business questions thomsonreuters.com ↗
  5. 2026 AI in Professional Services Report | Thomson Reuters Institute thomsonreuters.com ↗
  6. AI is Ready but Firms are Not: How Falling Behind on AI Implementation is Costing Clients and Talent prnewswire.com ↗
  7. Future of Professionals Report 2026 (PDF) datocms-assets.com ↗
  8. ILTA 2025 Technology Survey Results Released Today: Legal Technology Trends ediscoverytoday.com ↗
  9. Law Firms Embrace AI, But Full Deployment Remains Rare law360.com ↗
  10. The Legal Industry Report 2025 americanbar.org ↗
  11. Harbor 2025 Law Department Survey Reveals Surge in AI Integration, Falling Outside Counsel Spend harborglobal.com ↗
  12. CLOC Releases 2026 State of the Industry Report cloc.org ↗
  13. Core 12 - Evaluate the Maturity of your Legal Operations cloc.org ↗
  14. 2025 LDO Index: Legal departments want better service enhancement, but success metrics don’t always reflect priorities thomsonreuters.com ↗
  15. 2026 ACC Chief Legal Officers Survey KEY FINDINGS acc.com ↗
  16. Law Department Compensation Survey (2025 Executive Summary) acc.com ↗
  17. Knowledge & Innovation Counsel (New York) — employer career page mwe.wd5.myworkdayjobs.com ↗
  18. Director Knowledge Management & Innovation (New York) — employer career page pillsburylaw.wd5.myworkdayjobs.com ↗
  19. Legal Knowledge Engineer lw.com ↗
  20. Knowledge Management (KM) Legal AI Specialist — employer career page job-boards.greenhouse.io ↗
  21. Harvey Raises at $11 Billion Valuation to Scale Agents Across Law Firms and Enterprises harvey.ai ↗
  22. Top 10 GenAI Transparency & Readiness Questions for Outside Counsel acc.com ↗
  23. Formal Opinion 2024-5: Generative AI in the Practice of Law nycbar.org ↗
  24. ABA Formal Ethics Opinions Archive by Issue Dates americanbar.org ↗
  25. Policy Paper on the Use of AI Tools by Legal Professionals in Emerging Companies and Venture Capital nycbar.org ↗
  26. Mata v. Avianca — Opinion and Order on Sanctions storage.courtlistener.com ↗
  27. New York (statewide) 22 NYCRR Part 161 tracker legalaigovernance.com ↗
  28. Insurance Circular Letter No. 7 (2024) dfs.ny.gov ↗
  29. New York Department of Financial Services provides AI cybersecurity guidance: what you need to know reuters.com ↗
  30. New York joins California in regulating development of frontier AI models davispolk.com ↗
  31. NY Overhauls Transparency and Governance Requirements for Frontier AI Developers dwt.com ↗
  32. Automated Employment Decision Tools (AEDT) - DCWP nyc.gov ↗
  33. FTC Announces Crackdown on Deceptive AI Claims and Schemes ftc.gov ↗
  34. FTC announces new enforcement initiative targeting deceptive AI practices davispolk.com ↗
  35. AI Act | Shaping Europe’s digital future digital-strategy.ec.europa.eu ↗
  36. Number of US lawyers increases for first time since 2020, ABA report says abajournal.com ↗
  37. Table C. U.S. District Courts—Civil Cases Commenced, Terminated, and Pending … September 30, 2024 and 2025 uscourts.gov ↗
  38. Rising AI expenses could hinder BigLaw performance, new survey shows abajournal.com ↗

The 59 percent transparency gap and the 80 percent neither-require-nor-encourage figure are ACC/Everlaw, October 2025. Less-than-one-third awareness is Thomson Reuters Institute, February 2026. The 78/6 quality-versus-delivery split is Future of Professionals 2026. ILTA 2025 is a technology-leader sample of 580 firms, not a census of every New York panel. Harbor/CLOC is 135 departments, median revenue $13 billion. ACC compensation bands are national self-reports effective 1 March 2025. Part 161 is cited from the Legal AI Governance tracker quoting Administrative Order 75; the official UCS HTML could not be read. The October 2024 DFS Part 500 letter is cited from Reuters. Employer career-page salary bands are 2026 New York postings described by type and tier.

A quiet conversation

Scoring the stack — or staffing the seat that will?

We map in-house seats for general counsel and chief legal officers in New York when the brief is tool-governance, vendor management or a specialist AGC, and we are just as willing to say the logo is not a criterion as to open a search. Confidential, no obligation.