# The Principal Who Thinks IT Owns AI, While IT Thinks the Principals Do **By Dan Cumberland** · Published August 12, 2026 · Categories: AI Strategy ![Featured image for The Principal Who Thinks IT Owns AI, While IT Thinks the Principals Do](https://cms.dancumberlandlabs.com/uploads/featured_image_29f87ee093.webp) Illustration: Dan Cumberland Labs with Gemini. > 86% of firms have no one accountable for AI governance. Here's why principals and IT both assume the other has it — and the minimum viable structure to fix it. Your principals assume IT handles AI\. Your IT team assumes the principals are making the strategic calls\. Somewhere in that gap sit dozens of AI tools no one officially runs, a stack of stalled pilots, and a question nobody has answered: who owns this? Call it the AI governance vacuum\. It's not a technology problem— it's an accountability problem, and the data suggests it's the norm\. The Logicalis CIO Report 2026[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1) found that only 14% of companies have clearly defined who is responsible for AI governance at the leadership level\. The other 86% have departments, project managers, or individuals making AI decisions without anyone overseeing the whole picture\. This isn't random\. It's structural\. And the firms that close the ownership gap don't do it with a committee— they do it with one decision\. ## How the Vacuum Forms The gap exists because both sides are acting rationally under flawed assumptions\. IT inherits accountability through its technology mandate— AI looks like software, so it must be IT's\. Principals assume technology rollout decisions flow to IT, the way infrastructure always has\. Neither assumption gets examined out loud\. The data confirms the split\. According to [IBM](https://www.cio.com/article/4183249/cios-plagued-by-a-growing-ai-accountability-gap.html) research[2](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-2), two\-thirds of CIOs and CTOs report being accountable for AI systems they don't fully control\. Business units deploy tools— marketing teams connecting LLMs \(large language models\) to content workflows, finance running ChatGPT for forecasts, project teams accessing customer data through AI agents— and none of it surfaces to IT until something goes wrong\. On the principal's side, the picture is equally revealing\. A Dataiku survey[3](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-3) found that 70% of CEOs claim to be the primary driver of AI strategy\. But only 60% of those same CEOs participate in more than half of AI\-related decisions\. They've claimed the strategy without operationalizing it\. > "Ultimately, the CEO still owns the outcome and the CIO ends up managing the reality\."— Dataiku[3](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-3) As Ben Schein, Chief AI and Analytics Officer at Domo, put it: "The pace problem isn't usually that AI is being shipped recklessly; it's that AI is being adopted faster than governance frameworks can adapt\."[2](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-2) This pattern is predictable when accountability isn't made explicit\. The vacuum doesn't stay abstract— it shows up in your project failure rate and in every AI tool your team runs without IT's knowledge, and it's where [AI culture initiatives stall](/blog/building-ai-culture) before they start\. ## What AI Governance Accountability Actually Costs Undefined AI accountability has a measurable price\. The most direct: 48% of AI projects miss their business objectives when no one is explicitly responsible for outcomes across the leadership team, per the Logicalis CIO Report 2026[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1)\. The costs accumulate in four ways: - **Shadow AI proliferation**: 78% of employees already deploy AI tools without IT approval[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1)— this isn't rogue behavior, it's people filling a vacuum\. - **Orphaned agents**: According to Cynozure's 2026 State of Industry Report[4](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-4), the average enterprise runs 47 AI agents with no designated owner— autonomous systems with no human accountable for what they do, cost, or produce\. - **Contested ownership**: The VentureBeat Q2 2026 survey[5](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-5) found 17% of organizations have no one holding formal AI accountability at all; another 20% say ownership is unclear or contested between teams\. - **Maturity gap**: Firms with defined AI governance roles score 2\.6 on AI maturity versus 1\.8 without clear ownership— a meaningful gap on a four\-point scale, per analysis citing Cynozure data\.[12](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-12) And McKinsey's 2025 State of AI[6](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-6) found that only 28% of organizations have direct CEO oversight of AI governance\. That group shows the strongest correlation with bottom\-line AI performance\. The instinct when you see these numbers is to hand AI to IT and be done with it\. That instinct is wrong— and understanding why explains what the right answer actually is\. A sound [AI governance strategy](/blog/ai-governance-strategy) starts with the ownership question, not the tooling question\. ## IT's Lane and the Business's Lane IT should own the guardrails, not the roadmap\. Security, data governance, access controls, shadow AI detection, vendor compliance— that's IT's lane\. The business outcomes AI is supposed to improve? Those belong to the leader accountable for those outcomes\. ISACA is explicit: "Business leaders must retain accountability for how AI is used and how decisions are made\."[7](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-7) This isn't a preference— it's a governance principle\. AI outcomes are non\-delegable to technical specialists\. The mandate mismatch explains why IT\-default ownership fails\. IT's core function is risk elimination and infrastructure stability— the opposite of what effective AI adoption requires, which is rapid experimentation and tolerance for failure\.[8](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-8) The result when IT runs the show: adoption slows, pilots stall, and shadow AI proliferates anyway\. Here's the role split in operational terms: ```html-table
IT Owns (Guardrails)Business/Operations Owns (Roadmap)
Security and access controlsAI strategy and priorities
Data governance and complianceUse case selection
Shadow AI detectionOutcome accountability
Vendor risk evaluationAdoption decisions
Infrastructure and model accessROI measurement
Regulatory compliance infrastructureResource allocation
``` In practical terms: IT answers "how do we do this safely?" and the named executive answers "what should we do next?" The pen vs\. the plumbing is how this role split plays out in AEC and professional services\. IT holds the plumbing: the infrastructure, the guardrails, the security architecture\. Operations holds the pen: the strategy, the roadmap, the accountability for results\. [The breakdown by firm type and role is covered in detail here](/blog/construction-operations/)\. The Zweig Group frames it well for AEC firms specifically: AI isn't a tech problem, it's a trust problem\.[9](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-9) Successful adoption requires assigning ownership rather than dispersing responsibility\. The structure is clear\. The harder question for a 250\-person firm is: who specifically holds the pen? ## The Minimum Viable Structure for a 250\-Person Firm A 250\-person professional services firm doesn't need a Chief AI Officer\. It needs one executive's name attached to one accountability statement— and the organizational clarity that follows from that\. A full\-time CAIO runs $300,000–$500,000 annually\.[10](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-10) For a 250\-person firm, that's rarely justifiable\. A fractional model delivers equivalent strategic accountability at a fraction of that investment— typically in the $60,000–$180,000 per year range, depending on scope and engagement structure— but the structure matters more than the title\. Gartner research[11](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-11) shows that at large enterprises, 70% of Chief Data & Analytics Officers hold primary responsibility for AI strategy\. The principle scales down; the title and cost don't have to\. Here's the minimum viable structure: 1. **Named executive** \(typically COO or VP Operations\) with explicit AI accountability— owns the roadmap, sets adoption priorities, and serves as the internal escalation point for AI decisions\. 2. **IT as governance infrastructure owner**— guardrails, security, access controls, shadow AI detection\. IT answers the "how do we do this safely" question; the named executive answers the "what should we do next" question\. 3. **Fractional AI advisor** for strategic framework, tool evaluation, and governance architecture— [what a fractional AI officer actually does](/blog/what-is-a-fractional-ai-officer/) for firms that need the structure without the full\-time hire\. The critical principle: name individuals, not functions\. The moment you assign "IT" or "leadership" without a specific person's name, you're back in the vacuum\. Organizations with explicit governance roles score measurably higher on AI maturity— not because they have better tools, but because someone is accountable for using them well\. The structure is clear\. The question is whose name goes on the line\. ## The One Decision Name someone\. That's the decision\. Not a committee\. Not a policy task force\. One executive whose name is on the accountability line for AI outcomes in your firm\. That person is not responsible for building the technical infrastructure— IT owns that\. They're not responsible for becoming an AI engineer or solving every tool question\. They're responsible for the roadmap, the priorities, and the outcomes\. The firms that close the AI governance accountability gap don't do it with a policy\. They do it with one decision, clearly communicated, followed by one consistent question at every leadership meeting: is AI delivering on what we said it would? No matter the question, people are the answer\. AI governance is no different— and once ownership is clear, the more interesting questions can start\. If the ownership model itself is the hard part to design, that's what [AI strategy services](/services/ai-strategy) are for— firms that need the governance structure without building it from scratch\. ## FAQ ### Who should own AI strategy in a mid\-sized professional services firm? The executive accountable for the business outcomes AI is supposed to improve— typically the COO or VP Operations\. IT owns the governance guardrails: security, access controls, data governance, and shadow AI detection\. These two roles are complementary, not competing\. Separating them is the prerequisite for both to function\.[7](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-7) ### What happens when no one owns AI governance? 48% of AI projects miss their business objectives when responsibility is undefined across leadership teams, per the Logicalis CIO Report 2026\.[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1) Shadow AI proliferates: 78% of employees already deploy AI tools without IT approval, creating compliance and security exposure that surfaces only after a failure forces accountability\.[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1) ### Does a 250\-person firm need a Chief AI Officer? Not at the full\-time price\. A CAIO runs $300,000–$500,000 annually\.[10](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-10) Most mid\-market firms need a named executive with explicit AI accountability, supported by IT for governance infrastructure and optionally a fractional AI advisor for strategic framework and tool evaluation\. The title is a large\-enterprise construct; the accountability is universal\. ### Why shouldn't IT solely own AI strategy? IT's mandate is risk elimination and infrastructure stability— the opposite of what effective AI adoption requires\.[8](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-8) ISACA is explicit: business leaders cannot delegate accountability for AI decisions to technical specialists\.[7](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-7) IT should own the guardrails\. The business leader accountable for outcomes should own the roadmap\. ### How common is the AI governance ownership gap? Common— and the data is specific\. Only 14% of companies have clearly defined who is responsible for AI governance at the leadership level, per the Logicalis CIO Report 2026\.[1](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-1) The VentureBeat Q2 2026 survey[5](/blog/blog-the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do#ref-5) found that 17% of organizations have no one holding formal AI accountability at all\. ## References 1. Logicalis, "AI Governance 2026: Only 14% Have Clarified Who Is Responsible" \(2026\)— [https://www\.digital\-chiefs\.de/en/ai\-governance\-2026\-only\-14\-percent\-responsible/](https://www.digital-chiefs.de/en/ai-governance-2026-only-14-percent-responsible/) 2. IBM / CIO\.com, "CIOs Plagued by a Growing AI Accountability Gap" \(2025\)— [https://www\.cio\.com/article/4183249/cios\-plagued\-by\-a\-growing\-ai\-accountability\-gap\.html](https://www.cio.com/article/4183249/cios-plagued-by-a-growing-ai-accountability-gap.html) 3. Dataiku, "The AI Accountability Gap: CEOs Own Strategy, CIOs Carry Decisions" \(2025\)— [https://www\.dataiku\.com/stories/blog/ai\-accountability\-gap\-ceos](https://www.dataiku.com/stories/blog/ai-accountability-gap-ceos) 4. Cynozure / 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) 5. VentureBeat, "The Control Gap: Enterprise AI Organizations Have an Ownership Problem, Not a Technology Problem" \(2026\)— [https://venturebeat\.com/resources/the\-control\-gap\-enterprise\-ai\-organizations\-have\-an\-ownership\-problem\-not\-a\-technology\-problem\-and\-most\-are\-governing\-it\-by\-hand](https://venturebeat.com/resources/the-control-gap-enterprise-ai-organizations-have-an-ownership-problem-not-a-technology-problem-and-most-are-governing-it-by-hand) 6. McKinsey & Company, "The State of AI 2025: How Organizations Are Rewiring to Capture Value" \(2025\)— [https://www\.mckinsey\.com/~/media/mckinsey/business%20functions/quantumblack/our\-insights/the\-state\-of\-ai/2025/the\-state\-of\-ai\-how\-organizations\-are\-rewiring\-to\-capture\-value\_final\.pdf](https://www.mckinsey.com/~/media/mckinsey/business%20functions/quantumblack/our-insights/the-state-of-ai/2025/the-state-of-ai-how-organizations-are-rewiring-to-capture-value_final.pdf) 7. ISACA, "Responsible AI: From Emerging Technology to Executive Governance Imperative" \(2026\)— [https://www\.isaca\.org/resources/news\-and\-trends/isaca\-now\-blog/2026/responsible\-ai\-from\-emerging\-technology\-to\-executive\-governance\-imperative](https://www.isaca.org/resources/news-and-trends/isaca-now-blog/2026/responsible-ai-from-emerging-technology-to-executive-governance-imperative) 8. Sidecar\.ai, "Why Your IT Department Shouldn't 'Own' Your AI Strategy" \(2024\)— [https://sidecar\.ai/blog/association\-ai\-strategy\-it\-management\-governance](https://sidecar.ai/blog/association-ai-strategy-it-management-governance) 9. Zweig Group, "AI Governance and Trust in AEC Firms" \(2025\)— [https://zweiggroup\.com/blogs/the\-zweig\-letter/ai\-governance\-and\-trust\-in\-aec\-firms](https://zweiggroup.com/blogs/the-zweig-letter/ai-governance-and-trust-in-aec-firms) 10. Multiple \(CAIO role guides\), "Chief AI Officer \(CAIO\): Role Guide" \(2026\)— [https://www\.futureproofing\.dev/resources/enterprise\-ai\-talent\-strategy/chief\-ai\-officer\-role\-guide](https://www.futureproofing.dev/resources/enterprise-ai-talent-strategy/chief-ai-officer-role-guide) 11. Gartner, "Gartner Survey Finds 70% of CDAOs Are Responsible for AI Strategy and Operating Model" \(2025\)— [https://www\.gartner\.com/en/newsroom/press\-releases/2025\-05\-12\-gartner\-survey\-finds\-seventy\-percent\-of\-cdaos\-are\-responsible\-for\-artificial\-intelligence\-strategy\-and\-operating\-model](https://www.gartner.com/en/newsroom/press-releases/2025-05-12-gartner-survey-finds-seventy-percent-of-cdaos-are-responsible-for-artificial-intelligence-strategy-and-operating-model) 12. Rohit Prabhakar, "Who Owns AI Governance? CMO, CDO and CIO Roles in 2026" \(2026\)— [https://www\.rohitprabhakar\.com/blog/who\-owns\-ai\-governance/](https://www.rohitprabhakar.com/blog/who-owns-ai-governance/) --- **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/the-principal-who-thinks-it-owns-ai-while-it-thinks-the-principals-do/