← Findings 🕐 26 min read
Findings

Who Influences the C-Suite on AI: The Thought Leader Map

Brandon Sneider · March 2026

These are technical people who speak business. They bridge the gap between what AI can actually do and what executives are told it can do.

Executive Summary

  • Six distinct categories of influence shape how executives think about AI — technologist-translators, analyst firms, consulting partners, VC narratives, honest practitioners, and legal/governance voices. Each category reaches different C-suite titles through different channels.
  • Enterprises spend $100K-$500K+ annually on analyst research from Gartner and Forrester that directly shapes AI strategy. The analysts most executives can name by title but not by individual name wield enormous quiet power over procurement and roadmap decisions.
  • The most trusted voices share one trait: they report what fails, not just what works. Andrej Karpathy (1.9M X followers), Ethan Mollick (Wharton, 800K+ combined reach), and Simon Willison are disproportionately influential because they contradict vendor hype with evidence.
  • VC firms are the primary narrators of “what’s coming next.” Sequoia (Sonya Huang) and a16z (Martin Casado) define the categories that become board-level agenda items 12-18 months later. a16z spent $3.53M on federal lobbying in 2025 alone.
  • The Big Three consulting firms produce the surveys executives forward to their boards. McKinsey’s State of AI (1,993 respondents, 105 nations, 2025), BCG’s AI Radar (2,360 executives, 2026), and Deloitte’s State of AI in the Enterprise (3,235 leaders, 2025) are the documents that trigger budget conversations.
  • Legal AI is the fastest-moving governance category. 87% of general counsel now report AI use within their teams (up from 44% in 2025), and the EU AI Act (August 2026) and Colorado AI Act (June 2026) are forcing formalized AI policies from best practice to compliance obligation.

The Technologist-Translators

These are technical people who speak business. They bridge the gap between what AI can actually do and what executives are told it can do. Their credibility comes from having built production systems, not from selling products or advisory services.

Andrej Karpathy

Platform: X (1.9M followers), YouTube (500K+ subscribers), personal blog (karpathy.ai), Eureka Labs (his education company)

What he says that matters: In an October 2025 interview with Dwarkesh Patel, Karpathy stated AGI remains at least a decade away despite rapid LLM advances. His December 2025 “LLM Year in Review” post became the single most-shared technical assessment of the AI field that year. He frames AI advancement around “ghost intelligence” and “ambient programming” — terms that give executives a vocabulary beyond “chatbot.”

Credibility: HIGH. Former OpenAI co-founder, former Director of AI at Tesla. Does not sell enterprise software. His educational content (Eureka Labs) is his primary commercial activity. When Karpathy speaks, his audience includes both the engineers building AI systems and the executives funding them.

C-suite reach: CTOs, Chief AI Officers, technically-oriented CEOs. His content filters up because engineering teams share it internally.

Ethan Mollick

Platform: Substack “One Useful Thing” (estimated 300K-500K subscribers based on growth trajectory from 100K in late 2023), X (800K+ followers), Wharton co-director of Generative AI Labs

What he says that matters: Mollick frames AI as a coworker, not a tool — and backs it with academic rigor. His book Co-Intelligence was a New York Times bestseller and named best book of the year by The Economist and the Financial Times. In March 2026, he argued companies should treat AI as a system capable of carrying out hours of independent work, reshaping strategy, coding, and risk assessment. TIME named him one of the 100 Most Influential People in AI.

Credibility: HIGH. Tenured Wharton professor. No product to sell. His research is independent and published through academic and journalistic channels. He runs experiments with AI rather than speculating, and reports negative findings alongside positive ones.

C-suite reach: CEOs, Chief Strategy Officers, CHROs. Mollick is the person non-technical executives read to understand AI without feeling talked down to. His Wharton affiliation gives him automatic credibility in boardrooms.

Cassie Kozyrkov

Platform: Keynote speaking circuit (40+ countries, stages at the UN, WEF, Web Summit, SXSW), Medium, advisory roles with organizations including Gucci, NASA, Meta, Spotify, Salesforce, GSK

What she says that matters: Former Google Chief Decision Scientist who founded the field of Decision Intelligence. She argues against “AI-first” strategies in favor of decision-first strategies. In a 2025 Info-Tech interview, she debunked the idea that simply adopting AI creates value — the value comes from better decisions, and AI is one input. Her framing shifts executives from “how do we use AI?” to “what decisions are we trying to improve?”

Credibility: HIGH. Former Google senior leadership, now independent through her company Kozyr. No product dependency. Her advisory model means she has seen inside dozens of major organizations.

C-suite reach: CEOs, Chief Data Officers, Chief Strategy Officers. Her WEF and Davos presence means she reaches the global executive class directly.

Andrew Ng

Platform: The Batch newsletter (1.7M+ subscribers via DeepLearning.AI), Coursera (co-founder), AI Fund, LandingAI

What he says that matters: Ng operates as the patient educator of the executive class. In early 2026, he stated AGI remains decades away, tempering hype. Each issue of The Batch opens with a personal letter offering strategic commentary on AI developments. His position as both educator (Coursera, DeepLearning.AI) and investor (AI Fund) gives him dual perspective on what works in production versus what sounds good in pitch decks.

Credibility: MODERATE-HIGH. His educational platforms are genuinely valuable and widely used. However, he runs AI Fund (a venture studio) and LandingAI (an enterprise AI product), which creates commercial interests. His AI advice sometimes aligns with his investment thesis. Still, his track record and Stanford affiliation give him substantial credibility.

C-suite reach: Broad. The Batch at 1.7M subscribers reaches everyone from engineering managers to board directors. His Coursera courses have educated millions of professionals who now hold senior roles.

Simon Willison

Platform: Blog (simonwillison.net, 23 years running), Substack newsletter, X, RSS/Atom subscribers

What he says that matters: Willison is the practitioner’s practitioner. He coined the term “prompt injection” and has been documenting LLM capabilities and limitations in real-time with a level of detail no one else matches. In 2026, he launched a project documenting “Agentic Engineering Patterns” for tools like Claude Code and OpenAI Codex. Karpathy himself has cited Willison’s blog as a consistently high-quality resource.

Credibility: VERY HIGH. Fully independent. No VC funding, no enterprise product, no advisory firm. Django co-creator who pivoted to AI coverage because it was technically interesting. His influence is outsized relative to his follower count because the people who read him are the people building production systems.

C-suite reach: Indirect but high-leverage. CTOs and engineering VPs read him directly. His insights filter into engineering team recommendations that reach C-suite budget decisions. His influence on practitioners makes him a “second-order” influence on executive strategy.

Allie K. Miller

Platform: LinkedIn, Instagram, X, TikTok (2M+ combined followers), Open Machine advisory firm, AI-First Academy course (300K+ students)

What she says that matters: Miller is the most-followed voice in AI business on LinkedIn. Named to TIME100 Most Influential People in AI (2025) and ADWEEK’s AI Trailblazers Power 100. Her advisory firm Open Machine works with Novartis, CyberArk, ServiceNow, Warner Bros. Discovery, and a major state pension fund. She produces AI guidebooks focused on how to build, scale, and outperform with AI.

Credibility: MODERATE. Former Amazon (AWS) Global Head of ML for Startups. Her content is accessible and business-oriented, but her primary revenue comes from courses and advisory — which creates some alignment between her messaging and commercial interests. Less technically deep than Karpathy or Willison.

C-suite reach: CMOs, CHROs, and non-technical C-suite. Her LinkedIn dominance means she reaches the exact executives who are trying to understand AI without a technical background.

The Analyst Class

Analyst firms are where AI strategy becomes budget line items. CIOs and CTOs use Gartner and Forrester to justify decisions to their boards, evaluate vendors, and benchmark against peers. The individual analysts matter less than the institution — but they matter.

Gartner

Enterprise cost: $80K-$100K/year for advisory access (30-45 minute analyst calls). Entry-level read-only access runs $20K-$25K/year per user. Large enterprise agreements with premium services reach $500K-$2M+/year.

Key AI analysts (2025-2026):

  • Arun Chandrasekaran, Distinguished VP Analyst. Leads Gartner’s generative AI research, including the Hype Cycle for Generative AI and AI Predicts reports. Co-author of “Predicts 2026: AI’s Impact on the Future of Workforce.”
  • Svetlana Sicular, VP Analyst. Pioneered Gartner’s AI governance research over a decade ago. One of the first analysts ever covering AI. Presented the AI governance playbook at Gartner IT Symposium/Xpo.
  • Rita Sallam, Distinguished VP Analyst. Focuses on the intersection of data, AI, and human-machine collaboration.

What they say that matters: Gartner predicts 40% of enterprise apps will feature task-specific AI agents by 2026, up from less than 5% in 2025. Their Hype Cycle remains the single most-referenced framework in board presentations about AI readiness. Gartner’s March 2026 data and analytics predictions directly shape how enterprises plan AI budgets for the next fiscal year.

Credibility: MODERATE. Gartner’s business model — where vendors pay to participate in Magic Quadrants and enterprises pay for research — creates structural conflicts. Gartner research is directionally useful but should never be the sole input for a major AI decision. Their sample sizes are generally strong (thousands of respondents) but their framing sometimes favors vendor categories over practical reality.

C-suite reach: CIOs and CTOs primarily. Gartner is the default language of IT procurement. If a CIO needs to justify an AI platform purchase, a Gartner Magic Quadrant placement is often the opening slide.

Forrester

Enterprise cost: $25K-$75K/year for team-level subscriptions. Enterprise packages reach $50K-$300K+ depending on scope.

Key AI analysts (2025-2026):

  • Rowan Curran, Principal Analyst. Co-authors Forrester’s AI/ML Platform Wave and the “AI Platforms Landscape, Q1 2026.”
  • Mike Gualtieri, VP and Principal Analyst. 30+ years designing mission-critical applications. Co-authors the AI/ML Platform Wave alongside Curran.

What they say that matters: Forrester’s 2026 AI predictions took a notably bearish position: enterprises will defer 25% of planned AI spend into 2027 as financial rigor eliminates weak use cases. Fewer than one-third of decision-makers can tie AI value to financial growth. CFOs will increasingly gate AI investments based on demonstrated ROI. Their State of AI Survey (2025, 1,400+ global AI decision-makers) provides one of the better independent benchmarks on adoption reality versus hype.

Credibility: MODERATE-HIGH. Forrester’s 2026 position — that AI hype is outpacing delivery — is a contrarian call among analyst firms. This willingness to say “the emperor has fewer clothes than advertised” increases their credibility with skeptical executives. Their financial-discipline framing resonates with CFOs, who are increasingly part of AI purchase decisions.

C-suite reach: CIOs, CTOs, and increasingly CFOs. Forrester’s emphasis on ROI measurement makes their research useful in budget conversations where the CFO is the audience, not just the CIO.

IDC

Key AI analyst:

  • Ritu Jyoti, Group VP, Worldwide AI and Automation Research. Leads IDC’s AI research team and shapes their spending forecasts.

What they say that matters: IDC forecasts global AI solutions spending at $307 billion in 2025, rising to $632 billion by 2028. AI infrastructure spending hit $82 billion in Q2 2025 alone (166% year-over-year increase). Agentic AI is projected to exceed 26% of worldwide IT spending ($1.3 trillion) by 2029. These numbers set the macro narrative that every consulting firm and vendor deck references.

Credibility: MODERATE. IDC’s spending forecasts are widely cited but represent projections, not actuals. Their numbers often appear in vendor marketing as implied validation. Useful for understanding market direction; less useful for making specific organizational decisions.

C-suite reach: CFOs and CEOs who need market-size context for board presentations. IDC numbers are the “total addressable market” of AI thought leadership — everyone cites them, few question the methodology.

The Consulting Voices

The Big Three produce the surveys and frameworks that get emailed to boards. Their influence is structural: a McKinsey partner’s recommendation in a board meeting carries weight no blog post can match.

McKinsey / QuantumBlack

Key voice: Lareina Yee, Senior Partner. Nearly 25 years advising on growth, technology, and transformation. Global head of McKinsey Alliances and a director of the McKinsey Global Institute’s technology research.

What they publish that matters: McKinsey’s annual “State of AI” survey (March 2025: 1,993 respondents across 105 nations, GDP-weighted; November 2025 update) is the most-cited enterprise AI research globally. Co-authored by Alex Singla, Alexander Sukharevsky, Lareina Yee, and the QuantumBlack team. Their research titled “The CEO’s Guide to Generative AI” lands directly in inbox-to-boardroom pipeline. The November 2025 update focused on agentic AI as the next frontier, with only 6% of organizations properly “rewiring” to capture AI value.

Credibility: MODERATE. McKinsey’s research methodology is solid (large samples, global scope). But McKinsey sells AI transformation engagements through QuantumBlack, and their research consistently frames AI as a massive opportunity requiring organizational rewiring — which is precisely what McKinsey sells. Their finding that only 6% of organizations are properly structured for AI value is both plausible and commercially convenient. Read their data, discount their recommendations by 30%.

C-suite reach: CEOs and boards directly. McKinsey is the default language of the boardroom. When a board member says “I read that only 6% of companies are capturing AI value,” they read it in McKinsey.

BCG / BCG X

Key voice: Sylvain Duranton, Global Leader of BCG X (tech build and design division).

What they publish that matters: BCG’s “AI Radar 2026” (2,360 executives across 16 markets, including 640 CEOs at $500M+ revenue companies) found that CEOs are personally taking the lead on AI decision-making and upskilling. Duranton’s key message: “The true competitive advantage lies with those CEOs who will reshape functions end-to-end and invent new products and services that drive growth.” The report found regional divergence — Eastern CEOs (India, China, Japan, Middle East) are highly confident in AI ROI; Western CEOs are more cautious.

Credibility: MODERATE. Same structural issue as McKinsey — BCG sells AI transformation through BCG X. Their 10-20-70 framework (10% algorithm, 20% technology, 70% organizational change) is a legitimate insight and a pitch for BCG’s organizational change practice simultaneously. Sample sizes are strong.

C-suite reach: CEOs primarily. The AI Radar’s CEO-specific framing and its focus on CEO-led strategy give BCG direct access to the top of the organizational chart.

Deloitte AI Institute

What they publish that matters: Deloitte’s “State of AI in the Enterprise” (2026 edition, 3,235 leaders across 24 countries, surveyed August-September 2025) is the largest of the Big Four surveys. Key finding: worker access to AI rose 50% in 2025, but only 34% of leaders report truly reimagining the business with AI. The number of companies with 40%+ projects in production is set to double in six months. Their conclusion: governance — not technology — is the difference between scaling and stalling.

Credibility: MODERATE. Same structural conflicts as McKinsey and BCG. Deloitte’s AI services practice creates commercial incentive to emphasize both opportunity and complexity. Their sample size (3,235) is the largest among consulting firms, which adds statistical weight.

C-suite reach: CIOs and CTOs. Deloitte’s enterprise-IT framing makes their research most useful for technology leaders, while McKinsey and BCG speak more directly to CEOs and boards.

Accenture

Key voices: Julie Sweet (CEO), Sen Ramani (Global Lead for Data and AI), Arnab Chakraborty (Chief Responsible AI Officer)

What they say that matters: Accenture CEO Julie Sweet stated in March 2026 that AI proficiency is mandatory for promotion at Accenture — making it one of the first major employers to formally tie career progression to AI adoption. Their Technology Vision 2025 (25th annual edition) frames AI-powered autonomy as the defining shift in enterprise digitization. Accenture generated $2.7B in AI-related revenue in 2024, making them the largest AI services provider by revenue among the consulting firms.

Credibility: LOW-MODERATE. Accenture is simultaneously the analyst (Technology Vision), the advisor, and the implementer. Their $2.7B in AI services revenue means every published finding supports their commercial pipeline. Use their data as a market-size indicator, not as independent analysis.

C-suite reach: CIOs and Chief Digital Officers. Accenture’s implementation-heavy positioning means their research reaches the executives responsible for deploying AI, not just strategizing about it.

The VC and Investor Voices

VC firms shape the “what’s coming” narrative 12-24 months before it reaches enterprise budget cycles. When Sequoia or a16z declares a category, it becomes a board-level question within a year.

Andreessen Horowitz (a16z)

Key voices: Martin Casado (General Partner, leads $1.25B+ infrastructure practice), Marc Andreessen (co-founder)

Platform: a16z blog, podcasts, events. Casado’s infrastructure practice is the largest AI-focused VC portfolio in the industry. a16z raised $15 billion across five new funds in January 2026. Total AUM: $46 billion (ranked first among VC firms globally as of July 2025).

What they say that matters: Casado compares the current AI moment to “1996” — years of building ahead before the real value emerges. His and Sarah Wang’s research on where value will accrue in AI (application layer vs. infrastructure) directly shapes how enterprise buyers think about build-vs-buy decisions. a16z’s political influence is unprecedented for a VC firm: $3.53M in federal lobbying in 2025 (double their 2024 total), with senior White House officials and top congressional staffers now calling a16z first on AI policy measures.

Credibility: LOW-MODERATE for enterprise strategy advice. a16z’s portfolio includes many AI companies, which means their “what’s coming” narrative also serves as marketing for their portfolio. Their lobbying expenditures reveal an agenda beyond disinterested analysis. Their market data is useful; their strategic recommendations should be filtered through “who benefits?”

C-suite reach: CEOs and boards at the strategic level. a16z shapes the macro narrative about AI’s trajectory that filters into board conversations. When a board member asks “what’s the next wave?”, they’re often channeling a16z’s latest thesis.

Sequoia Capital

Key voices: Sonya Huang (General Partner, led investments in OpenAI, Hugging Face, Langchain, Glean), Pat Grady (Managing Partner)

Platform: Sequoia blog, “Training Data” podcast, annual AI Ascent conference, X (Huang: 17.3K followers)

What they say that matters: In January 2026, Huang and Grady argued we have entered the era of functional AGI — meaning AI systems can “figure things out” autonomously in ways that matter for real work. Huang’s vertical agents thesis predicts specialized AI handling legal research, medical documentation, and financial analysis will generate multi-trillion-dollar value. At AI Ascent 2025, Huang showed ChatGPT engagement approaching Reddit-level metrics. Sequoia’s 2026 prediction: “A Tale of Two AIs” — data center buildout delays alongside continued AI adoption growth.

Credibility: LOW-MODERATE. Same structural issue as a16z — Sequoia’s investments in OpenAI, Anthropic (January 2026 investment), and application-layer AI companies mean their thesis about where value accrues is also a portfolio-promotion exercise. Huang’s analysis is consistently high-quality and data-grounded; just remember who benefits from her conclusions.

C-suite reach: CEOs, Chief Strategy Officers, and board members. Sequoia’s “generative AI’s act two” framing and their AGI timeline predictions are the kind of content that gets forwarded in executive Slack channels.

Tomasz Tunguz / Theory Ventures

Platform: Blog (tomtunguz.com, 150K+ founders and operators, millions of pageviews monthly)

What he says that matters: Tunguz publishes data-driven annual predictions and scores himself publicly (7.85/10 on his 2025 predictions). His 2026 predictions cover AI agents, IPO timing, and infrastructure shifts. His Google background (product manager on AdSense, billion-dollar business unit) gives him operational credibility that most VCs lack. Theory Ventures focuses on data and data infrastructure, with eight unicorn investments including Looker, Monte Carlo, and Dremio.

Credibility: MODERATE. He scores his own predictions publicly and honestly — a rare practice in VC thought leadership. His investment thesis in data infrastructure creates some commercial bias, but his willingness to be graded on accuracy increases trust.

C-suite reach: CTOs, Chief Data Officers, and VP Engineering. Tunguz’s content is more operational than strategic, making it useful for technology leaders rather than boards. His self-scoring creates accountability that other VC voices lack.

The Practitioner Voices

These are the people building real AI systems and publishing honest accounts of what works and what fails. They have no product to sell and no fund to raise. Their influence comes from credibility born of operational reality.

Gergely Orosz / The Pragmatic Engineer

Platform: Substack (#1 technology newsletter on the platform), podcast (10M+ downloads by end of 2025), YouTube. Pragmatic Summit (February 2026, 400 top engineers)

What he says that matters: Orosz documents how AI is changing software engineering with uncomfortable specificity. His coverage of AI coding tools in 2025-2026 tracks adoption across Copilot, Cursor, Claude Code, and MCP protocol — noting that agents are “slowly overtaking” autocomplete as the primary AI coding pattern. He interviews engineering leaders at major companies about their actual AI deployment results, not their marketing claims.

Credibility: VERY HIGH. Former Uber and Microsoft engineer. No VC backing, no product to sell. His newsletter is subscription-supported, meaning his incentive is reader trust. His coverage of METR’s finding that experienced developers are 19% slower with AI tools is the kind of data that vendor-aligned publications suppress.

C-suite reach: VP Engineering, CTOs, and engineering-minded CEOs. His Pragmatic Summit drew 400 senior engineering leaders, creating a direct channel to technology decision-makers. His newsletter shapes how engineering leaders advise their executives on AI tool investments.

Chip Huyen

Platform: Personal blog (huyenchip.com), books, GitHub, LinkedIn (Top Voice in Data Science & AI)

What she says that matters: Her book AI Engineering (2025) is the most-read book on the O’Reilly platform since its launch. Her earlier Designing Machine Learning Systems (2022) is an Amazon bestseller translated into 10+ languages. She focuses on the production gap — the difference between a working demo and a reliable system. Her MLOps guide and AI engineering curriculum are used by teams at companies from startups to FAANG.

Credibility: VERY HIGH. Former Stanford instructor, worked at Snorkel AI, NVIDIA, and started an AI infrastructure company. Her content is rigorously practical. LinkedIn recognized her as a Top Voice in Software Development (2019) and Data Science & AI (2020).

C-suite reach: Indirect. CTOs and VP Engineering use her frameworks to evaluate AI infrastructure decisions. Her books shape how engineering teams think about production readiness, which filters into timeline and budget conversations with executives.

Swyx (Shawn Wang) / Latent Space

Platform: Latent Space podcast and Substack (10M+ readers and listeners in 2025, up from 2M in 2024), AI Engineer conference series (7 events planned for 2026 globally)

What he says that matters: Latent Space covers the technical frontier of AI engineering — the specifics of building with LLMs, agents, and multimodal systems. His AI Engineer conference series (growing from 1 event/year in 2023 to 7 in 2026) has become the gathering point for practitioners who build AI products as opposed to those who talk about them. His “Scaling Without Slop” thesis for 2026 addresses quality degradation as AI-generated content floods the internet.

Credibility: HIGH. Practitioner background (DX, Netlify). No VC fund. His conference and newsletter are his business, but they’re built on trust with a highly technical audience that would detect hype instantly.

C-suite reach: CTOs and Chief AI Officers at companies building AI products. His audience skews more technical than Orosz’s, but his conference series provides direct access to AI engineering leadership at scale.

Google Office of the CTO / Will Grannis

What they say that matters: Google Cloud CTO Will Grannis published a candid 2025 retrospective acknowledging that even his own team struggled to build a reliable internal AI agent — it only worked after leadership forced process discipline. His key insight: “AI exposes hidden process chaos. Organizations must tighten their operating discipline rather than simply bolt AI onto old habits.” This is one of the few honest accounts from a hyperscaler about the organizational difficulty of AI deployment.

Credibility: MODERATE. Grannis works for Google Cloud, which sells AI infrastructure. But his willingness to publish failure stories and organizational challenges is rare among cloud providers and increases trust.

C-suite reach: CIOs evaluating cloud AI platforms. Grannis’s CXOTalk appearances and Google Cloud blog posts reach technology leaders directly.

Matt Turck / FirstMark (MAD Landscape)

Platform: Annual MAD (Machine Learning, AI & Data) Landscape map, MAD Podcast, Data Driven NYC

What he says that matters: The annual MAD Landscape is the most widely shared visualization of the AI/ML/data ecosystem. The 2025 edition framed the market as simultaneously “Bubbling and Building” — valuations are often stratospheric with a clear AI premium, but real products are shipping underneath the hype. Available in both PDF and interactive searchable format at mad.firstmark.com.

Credibility: MODERATE. Turck runs FirstMark, a VC firm, so his landscape includes portfolio companies. But the comprehensiveness of the map — covering the entire ecosystem — makes it a reference tool rather than marketing.

C-suite reach: CTOs and Chief Data Officers use the MAD Landscape to understand vendor categories and competitive dynamics. Board members occasionally see it as context in technology strategy presentations.

Benedict Evans

Platform: Semi-annual presentations (“AI Eats the World”), personal blog (ben-evans.com), keynotes at Slush, SuperAI, B2BMX

What he says that matters: Evans produces the most widely shared macro-level analysis of AI’s impact on the tech industry. His presentation decks circulate among executives like research reports. The November 2025 “AI Eats the World” covered $400B in AI platform investments, common misconceptions about AI, and integration hurdles. He presents to corporate audiences including Alphabet, Amazon, AT&T, Axa, LVMH, Nasdaq, Swiss Re, and Vodafone.

Credibility: HIGH. Former a16z partner, now independent. His VC background gives him structural understanding of market dynamics. His independence since leaving a16z means no portfolio to promote. His presentations are free and publicly available, which builds trust.

C-suite reach: CEOs and boards directly. His corporate speaking roster (L’Oreal, LVMH, Deutsche Telekom, Vodafone) puts him in front of European and global C-suites. His slide decks are the kind of content that boards receive as pre-meeting reading.

General counsel are the fastest-growing AI constituency in the C-suite. With the EU AI Act (August 2026), Colorado AI Act (June 2026), and proliferating state requirements, legal AI governance has moved from “interesting” to “mandatory.”

The Numbers

AI adoption in corporate legal departments nearly doubled year over year: 87% of general counsel now report AI use within their teams, compared with 44% in 2025 (The General Counsel Report, March 2026). Legal departments with formalized technology roadmaps hit an all-time high of 53%, more than double from 25% the prior year. Legal tech spending surged 9.7% as firms race to integrate AI.

Bob Ambrogi — Lawyer and journalist. Writes LawSites (now LawNext), the award-winning blog covering legal technology. Hosts the LawNext podcast and Legaltech Week roundtable. Named among Clio’s top legal influencers. His coverage of products, ethics, and business models shapes early-adopter perception of new legal AI tools. When a legal tech product launches, Ambrogi’s review determines its initial reception among forward-looking firms.

Casey Flaherty — Co-founder and Chief Strategy Officer at LexFusion (now Baretz & Brunelle). Helps firms and corporate legal departments define and measure innovation. His work directly influences how major legal buyers design technology stacks and evaluate vendors. Flaherty is a pragmatic voice who focuses on measurement and accountability rather than hype.

Daniel W. Linna Jr. — Director of Law and Technology Initiatives at Northwestern Pritzker School of Law and McCormick School of Engineering. Selected to serve on the Illinois Supreme Court AI Task Force (2025). Member of IEEE P2863 Working Group developing AI governance standards. Founder of LegalRnD at Michigan State. His academic rigor and multi-disciplinary position (law + engineering) give him credibility with both legal practitioners and technologists.

Jessica Lee — Chair of Privacy, Security & Data Innovations at Loeb & Loeb. Named to the 2026 Lawdragon 100 Leading AI & Legal Tech Advisors. An early thought leader on AI’s impact on data security and privacy. Works with some of the world’s largest companies to build AI governance structures.

Jay Edelson — CEO and founder of Edelson PC. Secured a landmark $1.5B settlement against Anthropic over the use of copyrighted materials to train generative AI models. Currently handling a case against OpenAI involving a teen user fatality. His litigation directly shapes the risk calculus for every company deploying AI.

AI Ethics and Governance Voices

Timnit Gebru — Founder of the Distributed AI Research Institute (DAIR). Former co-lead of Google’s Ethical AI team. Co-authored foundational research on racial and gender bias in facial recognition. Her work shapes AI governance regulations worldwide. Her controversial exit from Google in 2020 made AI ethics a permanent board-level topic.

Joy Buolamwini — Founder of the Algorithmic Justice League. Co-authored the influential “Gender Shades” paper with Gebru. Her research on bias in commercial AI systems led directly to corporate policy changes at IBM and Microsoft. Her advocacy frames AI governance as a civil rights issue, which resonates with general counsel navigating discrimination liability.

Credibility note on ethics voices: Gebru and Buolamwini are fully independent researchers with no commercial AI products. Their credibility with legal and governance teams is high precisely because they have been willing to challenge the largest AI companies. Some technology executives dismiss them as critics rather than contributors — which is a mistake, because regulators and courts take their research seriously.

Key Data Points

Voice / Entity Platform Reach Credibility Primary C-Suite Audience
Andrej Karpathy X, YouTube, Blog 1.9M (X), 500K+ (YouTube) HIGH (independent, technical) CTOs, Chief AI Officers
Ethan Mollick Substack, X 300K-500K (Substack), 800K+ (X) HIGH (academic, independent) CEOs, CSOs, CHROs
Andrew Ng Newsletter, Coursera 1.7M (The Batch) MODERATE-HIGH (also investor) Broad C-suite
Cassie Kozyrkov Speaking, advisory Global executive circuit HIGH (independent, ex-Google) CEOs, CDOs
Simon Willison Blog, Substack Smaller but high-leverage VERY HIGH (fully independent) CTOs via engineers
Allie K. Miller LinkedIn, social 2M+ combined MODERATE (advisory + course revenue) CMOs, CHROs, non-tech C-suite
Gartner Research platform Dominant in enterprise IT MODERATE (vendor-funded model) CIOs, CTOs
Forrester Research platform Strong in enterprise IT MODERATE-HIGH (bearish, independent) CIOs, CTOs, CFOs
McKinsey / QuantumBlack Reports, events Global reach (1,993-respondent survey) MODERATE (also sells AI services) CEOs, boards
BCG / BCG X Reports, events 2,360-exec survey MODERATE (also sells AI services) CEOs
Deloitte AI Institute Reports 3,235-leader survey MODERATE (also sells AI services) CIOs, CTOs
a16z Blog, events, lobbying $46B AUM, policy influence LOW-MODERATE (portfolio bias) CEOs, boards
Sequoia Blog, podcast, events OpenAI/Anthropic investor LOW-MODERATE (portfolio bias) CEOs, CSOs
Gergely Orosz Substack, podcast #1 tech newsletter, 10M+ downloads VERY HIGH (independent) VP Eng, CTOs
Chip Huyen Books, blog #1 O’Reilly AI book VERY HIGH (independent) CTOs, VP Eng
Swyx / Latent Space Podcast, events 10M+ readers/listeners HIGH (practitioner) CTOs, Chief AI Officers
Bob Ambrogi LawNext blog, podcast Legal tech community HIGH (journalist, independent) General Counsel
Timnit Gebru DAIR, research Academic and regulatory HIGH (independent researcher) General Counsel, boards

What This Means for Your Organization

The influence map reveals three dynamics that matter for executive AI strategy.

First, the C-suite is getting its AI information from sources with structural conflicts. The consulting firms that produce the surveys forwarded to boards (McKinsey, BCG, Deloitte) also sell AI transformation services. The VC firms that define “what’s next” (a16z, Sequoia) profit when enterprises adopt the categories they name. The analyst firms that evaluate vendors (Gartner, Forrester) are funded in part by those vendors. None of this makes their research worthless — it makes it essential to triangulate across sources rather than rely on any single one.

Second, the most credible voices are the least commercially motivated. Karpathy, Mollick, Willison, Orosz, and Huyen have no AI platform to sell. Their incentive is accuracy, because their audience (technical practitioners and informed executives) would detect and punish hype. When these voices say something works, it probably works. When they say something fails, the consulting firms’ surveys will confirm it 18 months later.

Third, legal and governance voices are underweighted in most AI strategy conversations. General counsel adoption of AI doubled in a single year. The EU AI Act takes effect in August 2026. Companies that treat AI governance as a legal compliance exercise rather than a strategic function are building on unstable ground. The voices in the legal/governance category (Ambrogi, Flaherty, Linna, Gebru) should be on the reading list alongside the technology voices.

The practical implication: build your AI intelligence function around independent voices for honest assessment, consulting surveys for peer benchmarking, and legal/governance voices for risk framing. Use VC narratives as leading indicators of what vendors will pitch you next year, not as strategic guidance.

If you are trying to separate signal from noise in the AI landscape and want to pressure-test your strategy against what the evidence actually supports, that is a conversation I am always happy to have.

Sources

  • McKinsey QuantumBlack, “The State of AI” (March 2025, n=1,993 across 105 nations; November 2025 update). mckinsey.com
  • BCG, “AI Radar 2026: As AI Investments Surge, CEOs Take the Lead” (February 2026, n=2,360 executives including 640 CEOs). prnewswire.com
  • Deloitte AI Institute, “State of AI in the Enterprise 2026” (n=3,235 leaders, 24 countries, August-September 2025). deloitte.com
  • Forrester, “Predictions 2026: Artificial Intelligence” (October 2025). forrester.com
  • Forrester, “State of AI Survey 2025” (n=1,400+ global AI decision-makers). forrester.com
  • Gartner, “Top Predictions for Data and Analytics in 2026” (March 2026). gartner.com
  • IDC, AI infrastructure spending forecast ($758B by 2029). idc.com
  • The General Counsel Report, “AI Adoption in Corporate Legal Departments Doubles” (March 2026). globenewswire.com
  • Fortune, “Did an OpenAI cofounder just pop the AI bubble?” — Karpathy interview (October 2025). fortune.com
  • Ethan Mollick, “One Useful Thing” Substack. oneusefulthing.org
  • Simon Willison, “Agentic Engineering Patterns” (2026). simonw.substack.com
  • LawFuel, “15 Legal Tech Leaders Shaping AI for Law in 2026.” lawfuel.com
  • Lawdragon, “The 2026 Lawdragon 100 Leading AI & Legal Tech Advisors.” lawdragon.com
  • a16z lobbying spend and policy influence. techflowpost.com
  • Sequoia Capital, “AI in 2026: A Tale of Two AIs.” sequoiacap.com
  • Latent Space 2026 plans. latent.space
  • Matt Turck / FirstMark, “Bubble & Build: The 2025 MAD Landscape.” mattturck.com
  • Chip Huyen, AI Engineering (2025, O’Reilly). huyenchip.com
  • Google Cloud Office of the CTO, “Lessons from 2025 on agents and trust.” cloud.google.com
  • Gergely Orosz, The Pragmatic Engineer. pragmaticengineer.com
  • Benedict Evans, “AI Eats the World” (November 2025). ben-evans.com
  • TIME100 Most Influential People in AI (2025). time.com
  • Vendr, Gartner pricing data. vendr.com

Credibility note: All consulting firm surveys (McKinsey, BCG, Deloitte, Accenture) are produced by organizations that sell AI services. Their data is useful for benchmarking; their recommendations should be filtered through commercial interest. Independent voices (Karpathy, Mollick, Willison, Orosz, Huyen, Ambrogi) have no product to sell and report both successes and failures. VC research (a16z, Sequoia, Tunguz) is directionally useful for identifying emerging categories but reflects portfolio incentives. Analyst research (Gartner, Forrester, IDC) is the default language of enterprise procurement; Forrester’s 2026 bearish position on AI hype adds useful counterweight.

Brandon Sneider | brandon@brandonsneider.com March 2026