# Nobody Owns AI (And That's Why Nothing Ships) **By Dan Cumberland** · Published August 7, 2026 · Categories: AI Strategy ![Featured image for Nobody Owns AI (And That's Why Nothing Ships)](https://cms.dancumberlandlabs.com/uploads/featured_image_13a0a6f06b.webp) Illustration: Dan Cumberland Labs with Gemini. > The pattern plays out at professional services firms constantly. A leadership team— four or five people around the table, all sharp, all using AI in some... The pattern plays out at professional services firms constantly\. A leadership team— four or five people around the table, all sharp, all using AI in some form— gathers to decide where to take this next\. Someone asks: who owns AI in your organization? The CIO says it's an operational call\. The COO says the technical infrastructure is IT's problem\. The CFO wants an ROI projection before anyone commits\. The business unit leads mention they're already doing their own thing on the side\. Nobody killed the initiative\. Nobody owns it, either\. The pilot that worked in Q1 is still sitting in pilot in Q4\. This isn't a technology problem\. The tools work\. It's an accountability problem— and the frustration is universal\. The data bears it out\. ## How Bad Is It, Really? More than 80% of AI projects fail to deliver intended business value— twice the failure rate of non\-AI IT projects— according to RAND Corporation research based on interviews with 65 experienced data scientists[1](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-1)\. The failure breakdown is precise: 33\.8% of AI projects were abandoned before reaching production\. Another 28\.4% were completed but delivered no value\. And 18\.1% delivered insufficient value to justify the investment\. RAND's conclusion is direct: "The root causes are overwhelmingly organizational and process\-oriented, not technical\."[1](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-1) The pattern in this data isn't ambiguous: the tools aren't the variable\. The same AI models available to firms that succeed are available to firms that fail\. The gap is organizational\. The abandonment rate is accelerating\. S&P Global Market Intelligence found that 42% of companies abandoned most of their AI initiatives in 2025— more than double the 17% rate from 2024[2](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-2)\. One year\. More than double\. MIT's 2025 analysis of 300 public AI deployments and 52 executive interviews found that 95% of generative AI pilots delivered no measurable P&L impact[3](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-3)\. Ninety\-five percent\. Not a rounding error\. Three failure modes, every time: - **33\.8%** — projects abandoned before ever reaching production - **28\.4%** — completed but delivered no business value - **18\.1%** — delivered insufficient value to justify the cost The failure isn't in the models\. It's in the meeting rooms\. ## When Nobody Owns AI in Your Organization — The Root Cause The tech is the easy part\. The human change— specifically, deciding who is accountable for AI outcomes— is where organizations break down\. AI governance is the system of accountability that determines who can approve, modify, or retire AI tools within an organization— and in most mid\-sized firms, that system doesn't exist\. Our [guide to AI governance strategy](/blog/ai-governance-strategy) covers the structural options in depth, but the diagnosis comes first: McKinsey found only 9% of enterprises describe their AI governance as mature, and just 28% of CEOs take direct responsibility for AI governance oversight[4](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-4)\. Deloitte's more recent 2026 survey shows some improvement: 21% of organizations now report mature governance[5](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-5)\. Both numbers say the same thing— four out of five organizations are still operating without a real accountability structure in place\. The Cynozure 2026 data, reported by Larridin, is more direct: 17% of organizations have no clear AI owner at all, and 77% have AI adoption outpacing their governance controls[6](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-6)\. Governance failure takes a few specific shapes\. Recognizing which one is yours matters: - **No owner at all:** No person or team holds formal accountability for AI outcomes - **Shared ownership that defaults to no one:** Multiple stakeholders claim partial responsibility; decisions stall when action is needed - **Tool\-by\-tool management without an enterprise view:** Each team adopts independently; no one sees the whole picture DataGrail summarized it plainly: "No role today is purpose\-built to own AI governance\."[7](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-7) That's not an accident\. AI cuts across operations, legal, technology, finance, and client delivery simultaneously\. It doesn't fit cleanly into any existing lane\. So every executive defers to the others\. The dysfunction has a specific shape— and understanding it is the first step toward fixing it\. ## The Six\-Executive Problem \(And Why Adding a CAIO Often Makes It Worse\) When no one owns AI, it's rarely because everyone ignored the question\. More often, six different executives each have a legitimate claim— and none will yield to the others\. - **CIO** — owns technology infrastructure and systems - **COO** — owns operational efficiency and process - **CFO** — owns budget approval and ROI accountability - **CHRO** — owns people, policy, and change management - **Chief Risk Officer \(CRO\)** — owns client data and revenue\-facing risk - **CDO** — owns data strategy and governance Each rationale is valid\. That's the problem\. IBM Institute for Business Value found that two\-thirds of CIOs are accountable for AI systems they don't control— a structural contradiction that produces exactly this stalemate[8](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-8)\. The instinctive response: hire a Chief AI Officer\. IBM data shows the CAIO role grew from 26% to 76% of surveyed organizations in a single year, and organizations with a CAIO do see approximately 10% better ROI on AI spend[8](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-8)\. But Executive Resilience Insider's analysis makes the catch plain: a CAIO without defined authority doesn't resolve the battle— it adds a seventh claimant[9](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-9)\. One documented case: an insurance company spent six months and created an entirely new C\-suite role to resolve a governance conflict that structured ownership could have addressed in a week\. If you're evaluating [what a fractional AI officer actually does](/blog/what-is-a-fractional-ai-officer) versus naming an internal owner, the distinction is outcome authority— not the title on the door\. Ownership isn't a job description\. It's accountability\. Someone needs to be answerable for whether AI initiatives produce results, not just whether the tools are running\. Meanwhile, the problem doesn't wait for the org chart to sort itself out\. ## What Happens When No One Governs — Shadow AI When no one officially owns AI, employees make the decision themselves— using tools their IT team has never seen and their firm has never approved\. Shadow AI exists because AI governance does not— where oversight is absent, unsanctioned tool adoption fills the void\. Zluri's 2025 research found that 80% of enterprise AI tools operate without IT management, and IT and security teams have visibility into fewer than 20% of the AI apps their organizations actually use[10](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-10)\. > "80% of enterprise AI tools operate without IT oversight— affecting organizations with 100\+ AI apps in use\." \(Zluri, 2025\) IBM's 2025 Cost of Data Breach Report found that 1 in 5 organizations has experienced a data breach linked to unsanctioned AI use[10](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-10)\. For a professional services firm, the risk is specific: client data flowing through personal AI accounts, unapproved tools making client\-facing recommendations without anyone's knowledge or consent\. Shadow AI isn't rebellion\. It's employees trying to do their jobs\. The governance vacuum created the void; they're filling it\. The good news: fixing this doesn't require a Chief AI Officer\. It requires a decision\. ## The Minimum Viable Fix — Ownership Without Bureaucracy A 250\-person firm doesn't need a Chief AI Officer\. It needs one named owner per initiative, a tiered risk classification, and a weekly review cadence— and it can have all three in place by the end of the week\. Minimum viable ownership has three components: ```html-table
ComponentWhat It IsWho Holds It
Named outcome ownerOne person accountable for AI results — not technical stewardship, outcome accountabilityTypically the COO or an operations-facing principal
Tiered risk classificationThree-level framework (low / medium / high) aligned to oversight requirementsDefined by named owner, approved by leadership
Weekly review cadence30-minute standing sync: what's in use, what's producing value, what needs a decisionRun by named owner
``` **Named outcome owner:** In our work with professional services firms, the named owner is typically the COO or an operations\-facing principal\. They don't need a new title\. They need formal authority over what gets approved, what gets used, and whether it's working\. **Tiered risk classification:** Not every AI tool needs the same level of oversight\. Low\-risk \(personal productivity, internal writing assistance\): use it, report it quarterly\. Medium\-risk \(client\-facing automation, process\-level decisions\): requires principal approval before deployment\. High\-risk \(data handling, financial decisions\): requires sign\-off and regular review\. **Weekly review cadence:** A 30\-minute standing sync— what's in use, what's producing value, what needs a decision\. This is where intent becomes accountability\. Without the cadence, even a well\-designed ownership structure stays theoretical\. What this doesn't require: a new C\-suite hire, a dedicated AI team, a formal governance committee, or a months\-long strategy process\. One professional services firm in the healthcare billing space demonstrates what becomes possible once AI adoption gets structure behind it\. Working with a single AI implementation partner to unify their content voice and scale educational materials, they saw visibility increase 300%\+ and eliminated the internal friction that had previously stalled progress[11](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-11)\. The expertise was always there\. The structure and accountability unlocked it\. The expertise your people already have is the ingredient\. AI amplifies it\. But amplification requires someone running the kitchen— the sous chef can't cook the meal if no one's ordered the dish\. ## FAQ ### What percentage of AI projects fail? More than 80% of AI projects fail to deliver intended business value, according to RAND Corporation research published April 2026 and based on interviews with 65 experienced data scientists[1](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-1)\. A separate S&P Global Market Intelligence survey found that 42% of companies abandoned most of their AI initiatives in 2025, up from 17% in 2024[2](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-2)\. Both figures point in the same direction: AI initiative failure is common, and the rate is accelerating\. ### Who should own AI governance in a mid\-sized company? In a company of 50–250 people, AI governance should sit with a named executive owner— typically the COO or a senior operations leader— who holds outcome accountability, not just technical oversight\. This person doesn't need a new title; they need clear authority over which AI tools get approved, how risks get classified, and whether initiatives are producing results\. For the structural options, see our [guide to AI governance strategy](/blog/ai-governance-strategy)\. ### What is shadow AI and why does it happen? Shadow AI refers to AI tools employees use without IT knowledge or organizational approval\. It happens when governance is absent— employees need tools to do their jobs, and when no one has defined what's approved, they make the decision themselves\. Zluri's 2025 research found that 80% of enterprise AI tools operate without IT management[10](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-10)\. ### Do I need a Chief AI Officer to fix this? Most mid\-sized firms don't need a dedicated Chief AI Officer\. They need a named owner with outcome authority— an existing executive who is formally responsible for AI results\. The CAIO model can work at scale, but IBM Institute for Business Value research shows that without clear authority structure, adding a CAIO often creates a seventh claimant to ownership rather than resolving the conflict[8](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-8)\. For more on [what a fractional AI officer actually does](/blog/what-is-a-fractional-ai-officer) versus naming an internal lead, that's a useful comparison\. ### Why do AI pilots succeed but never ship to production? The pilot paradox describes the structural reality that AI pilots succeed under narrow conditions— small team, clean data, limited accountability— that disappear at production scale\. Cross\-functional sign\-offs are suddenly required, and no single person has the authority to approve deployment\. The governance gap that was invisible in the pilot becomes a blocking wall at production\. The fix is the same: name an owner before the pilot starts\. Industry data suggests the majority of successful AI pilots never reach full production deployment[12](/blog/blog-nobody-owns-ai-and-that-s-why-nothing-ships#ref-12)\. ## Conclusion AI isn't failing your firm because the technology isn't ready\. It's failing because nobody owns it\. The governance vacuum isn't a technology gap or a budget gap\. It's an accountability gap\. It doesn't require a new C\-suite hire to fix— it requires one person with authority and a clear charge\. Once someone owns AI in your organization, the expertise your people already have gets amplified\. No matter the question, people are the answer— and AI governance is no different\. If you're evaluating where to start, [building AI culture in your team](/blog/building-ai-culture) and using an [AI decision framework for founders](/blog/ai-decision-framework-founders) both give you structured next steps\. Both require a named owner before they go anywhere\. That's step one\. Everything else follows\. ## References 1. RAND Corporation, "The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed: Avoiding the Anti\-Patterns of AI" \(April 2026\) — [https://www\.rand\.org/pubs/research\_reports/RRA2680\-1\.html](https://www.rand.org/pubs/research_reports/RRA2680-1.html) 2. S&P Global Market Intelligence via CIO Dive, "AI project failure rates are on the rise" \(2025\) — [https://www\.ciodive\.com/news/AI\-project\-fail\-data\-SPGlobal/742590/](https://www.ciodive.com/news/AI-project-fail-data-SPGlobal/742590/) 3. MIT NANDA via Virtualization Review, "MIT Report Finds Most AI Business Investments Fail, Reveals 'GenAI Divide'" \(August 2025\) — [https://virtualizationreview\.com/articles/2025/08/19/mit\-report\-finds\-most\-ai\-business\-investments\-fail\-reveals\-genai\-divide\.aspx](https://virtualizationreview.com/articles/2025/08/19/mit-report-finds-most-ai-business-investments-fail-reveals-genai-divide.aspx) 4. McKinsey & Company, "The State of AI in Early 2024: Gen AI Adoption Spikes and Starts to Generate Value" \(2024\) — [https://www\.mckinsey\.com/capabilities/quantumblack/our\-insights/the\-state\-of\-ai\-2024](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-2024) 5. Deloitte, "The State of AI in the Enterprise — 2026 AI Report" \(2026\) — [https://www\.deloitte\.com/us/en/what\-we\-do/capabilities/applied\-artificial\-intelligence/content/state\-of\-ai\-in\-the\-enterprise\.html](https://www.deloitte.com/us/en/what-we-do/capabilities/applied-artificial-intelligence/content/state-of-ai-in-the-enterprise.html) 6. Larridin, "The AI Accountability Vacuum: When No One Owns Outcomes, Everyone Pays" \(2026\) — [https://larridin\.com/blog/ai\-accountability\-ownership](https://larridin.com/blog/ai-accountability-ownership) 7. DataGrail, "Who Owns AI Governance?" \(2025\) — [https://www\.datagrail\.io/blog/ai\-governance/who\-owns\-ai\-governance/](https://www.datagrail.io/blog/ai-governance/who-owns-ai-governance/) 8. IBM Institute for Business Value, "The Rise and ROI of the Chief AI Officer" \(2026\) — [https://www\.ibm\.com/think/news/rise\-chief\-ai\-officer](https://www.ibm.com/think/news/rise-chief-ai-officer) 9. Executive Resilience Insider, "When Six Executives Claim AI Ownership, Nobody Decides" \(2026\) — [https://executiveresilienceinsider\.com/p/when\-six\-executives\-claim\-ai\-ownership\-nobody\-decides](https://executiveresilienceinsider.com/p/when-six-executives-claim-ai-ownership-nobody-decides) 10. Zluri via BusinessWire, "Zluri Report Exposes 'Shadow AI' Epidemic: 80% of Enterprise AI Tools Operate Unmanaged" \(June 2025\) — [https://www\.businesswire\.com/news/home/20250618738949/en/Zluri\-Report\-Exposes\-Shadow\-AI\-Epidemic\-80\-of\-Enterprise\-AI\-Tools\-Operate\-Unmanaged](https://www.businesswire.com/news/home/20250618738949/en/Zluri-Report-Exposes-Shadow-AI-Epidemic-80-of-Enterprise-AI-Tools-Operate-Unmanaged) 11. Jeremy Zug, Partner, Practice Solutions — internal case study \(Angle 8: Professional Services Scaler\) 12. Astrafy, "Scaling AI from Pilot Purgatory: Why Only 33% Reach Production and How to Beat the Odds" \(2025\) — [https://astrafy\.io/the\-hub/blog/technical/scaling\-ai\-from\-pilot\-purgatory\-why\-only\-33\-reach\-production\-and\-how\-to\-beat\-the\-odds](https://astrafy.io/the-hub/blog/technical/scaling-ai-from-pilot-purgatory-why-only-33-reach-production-and-how-to-beat-the-odds) --- **How we made this article:** We use AI in our research and writing so our small team can share more of what we learn. We verify the sources and take responsibility for every article we publish. [Read how we use AI.](https://dancumberlandlabs.com/how-we-use-ai/) --- ## About the author **Dan Cumberland** — Founder, Dan Cumberland Labs Dan Cumberland helps engineering and construction firms see where they stand with AI and decide what to build first. He created Pacemark, the AI maturity model behind that work, from research on more than 300 companies. - Take the assessment: https://pacemark.ai/signal/assessment/?track=aec&utm_source=dcl-site&utm_medium=link&utm_campaign=pacemark-assessment - Book a call: https://book.dancumberland.com/ai-strategy --- Source: https://dancumberlandlabs.com/blog/nobody-owns-ai-and-that-s-why-nothing-ships/