Most AI Failures Don't Start With the Code. They Start With the Contract w/ Rory O'Keeffe | S2E15

Guest: Rory O'Keeffe — MEET RORY ► 🔹 Fractional General Counsel and former Accenture Director of Legal Services and former Partner at Matheson LLP, with 20+ years’ experience in commercial law, technology law, AI governance, data protection, cybersecurity, digital regulation and complex contracting. I support high‑growth and enterprise organisations across the UK, EU and US, helping them manage regulatory risk, commercial negotiations, digital transformation and AI compliance. 🔹I’ve led legal strategy for £2bn+ annual contract value portfolios and negotiated 150+ major technology, outsourcing, SaaS and digital services agreements valued £20m–£250m. My focus: accelerating deal velocity, improving legal processes, and reducing operational and regulatory risk. 🔹 As Founder of RMOK Legal, I deliver embedded Fractional GC services that provide 55% cost savings vs. external counsel and 35% faster contract cycles. I build scalable governance frameworks for AI, data, cloud, cybersecurity and digital operations, enabling confident, defensible decision‑making at pace. 🔹 Co‑author of the bestselling “AI Advantage”, offering practical guidance for boards and legal teams on AI strategy, risk and responsible deployment. 🔹 Host of the “Beyond The Fine Print®” podcast, providing actionable insight on AI regulation, data protection, in‑house leadership, digital risk and legal transformation. 🔹 Active in shaping the profession as a Trustee of The Solicitors’ Charity, Trustee at Activity Alliance, and member of the Society for Computers and Law (AI Committee), focusing on digital ethics, AI governance and professional resilience. Speaker at SCL, Data Protection Conference 2026 and more. MY CLIENTS ► In‑house counsel, boutique firms, founders and C‑suite leaders managing high workloads, complex commercial issues, regulatory uncertainty, distributed teams and digital risk. WHAT I OFFER ► 🎯 Fractional GC services: embedded, commercially pragmatic, fast‑moving legal leadership. 🎯 High‑value commercial and technology contracting: SaaS, cloud, outsourcing, digital services, data & AI. 🎯 Strategic advisory: Legal risk, governance, board reporting, legal operations, process optimisation. I support clients facing mission‑critical negotiations, capacity gaps and regulatory change. At RMOK Legal… We Look After It™. WHY WORK WITH ME? ► 🔍 Clear, pragmatic, business‑aligned advice 🌍 Deep enterprise experience (Accenture, Matheson) 🎧 Trusted voice in AI, technology law and risk LET'S CONNECT ► 📧 rory.okeeffe@rmoklegal.com 📞 https://calendly.com/rory-okeeffe-rmoklegal/discovery-call

In this episode, we break down why AI deployment is not just a tech decision, it is a legal, commercial, and risk decision. Rory O'Keeffe, Commercial & AI Lawyer, shares how most failures do not come from the code itself, but from the contracts, the assumptions, and the lack of clarity around data. Founders often rush into AI tools thinking they are buying software, when in reality, they are taking on a new layer of risk. If you cannot clearly explain where your data goes, you are not deploying AI, you are outsourcing control. We also dive into what founders consistently miss, from weak contract terms to misunderstanding how data is used, stored, and even reused to train models. Rory explains why things like anonymization, auditability, and explainability are not just legal language, they are critical safeguards. The conversation highlights real examples of how AI can go wrong, not because of bad intent, but because of poor inputs, lack of oversight, or blind trust in automation. It is a reminder that AI is a tool, not the decision-maker. This episode is also a practical playbook for founders navigating AI. From asking the right questions before signing contracts to planning your exit before you even start, the focus is on staying in control. The biggest takeaway is simple, AI will not replace you, but someone using AI better than you might. The advantage goes to those who understand both the technology and the responsibility that comes with it. Tune in and learn how to deploy AI with clarity, protect your business, and make better decisions in a fast-moving world. -- TIMESTAMPS: 00:00 Intro 01:05 Most AI failures start in the contract, not the code 02:10 Welcome to The Human Protocol 03:20 Rory O’Keefe – Irish lawyer, founder of RMLK Legal & Beyond the Fine Print 05:40 What founders misunderstand: AI vs automation vs robotics 08:20 Eyes wide open — rushing AI can cripple your company 10:50 Most common contract mistake: Click-through terms & data use clauses 13:40 “We will anonymize your data” — what that really means (and the risks) 17:10 GDPR, CCPA & global privacy laws every founder must know 20:30 Auditability & explainable AI — can they show where it went wrong? 24:10 Agentic AI horror story: Booking a hotel with voice cloning & credit cards 28:50 Power of prompting + why “how” matters more than ever 32:40 Data privacy & security in law firms — on-prem, private instances & hyperscalers 37:20 IP ownership: Who owns prompts, outputs & generated content? 42:10 Indemnification, liability & who pays if AI causes harm 47:30 Termination & exit clauses — avoid tech debt & lock-in 52:10 Human in the loop: Left-brain data vs right-brain critical thinking 56:40 Advice to younger Rory & every founder in the AI era

Conversation Summary

Rory O'Keeffe, a Commercial & AI Lawyer, explains why many AI deployment failures originate from legal oversights rather than technical issues. AI contracts are not just about acquiring software but involve complex legal relationships that necessitate understanding the risk profile. O'Keeffe highlights how founders often overlook critical elements in AI agreements, such as data use clauses, which can lead to unintended consequences. This conversation serves as both a cautionary tale and a practical guide for founders, emphasizing the necessity of clear contracts and informed deployment strategies to maintain control and protection for their business.

Misunderstanding AI vs. Automation Leads to Rushed Decisions

Business leaders often confuse AI with automation, leading to rushed decisions about deploying AI technologies. Rory O'Keeffe stresses that AI has been around for decades but has only recently become affordable and accessible, prompting hasty adoption due to competitive pressures. However, using AI without understanding the risks can cripple a company.

Contracts Define Data Ownership and Limit Risk

O'Keeffe underscores that many AI failures stem from insufficient contractual terms, especially concerning data use. Many AI tools rely on feeding data for machine learning improvements. Thus, it's crucial for contracts to clearly define data ownership, anonymization processes, and usage rights to protect the proprietary data and reduce exposure to legal risks.

Ensure Auditability and Explainability in AI Systems

AI systems must have the capability to trace decision-making processes to ensure accountability. O'Keeffe highlights the importance of explainable AI, where failures can be audited to discern where and why an AI system went wrong, thus maintaining trust and compliance with regulatory standards.

AI Won't Replace Humans, But It Will Change Roles

AI is a tool that offers increased accuracy and efficiency but requires human oversight to guide and interpret its outputs. O'Keeffe argues that AI will not make roles like lawyers redundant; instead, those who adeptly use AI in their roles will replace those who do not, illustrating a shift in skill requirements rather than a wholesale replacement of human jobs.

Anticipate and Plan for AI Exit Strategies

Exit strategies in AI contracts are often neglected but critical. O'Keeffe advises on planning for termination of AI relationships by considering data transition, retention obligations, and ensuring that systems can be decoupled without accruing technical debt, thus maintaining business continuity after contract termination.

Final Thoughts

AI deployments in business are complex intersections of technology, legality, and strategy. Rory O'Keeffe emphasizes the vital role that contracts play in shaping successful and secure AI implementations. Businesses must navigate these complexities thoughtfully, ensuring that human oversight remains integral to managing AI systems. Understanding and negotiating the nuances of AI contracts can safeguard data, enhance innovation, and ensure that technology serves business objectives responsibly and effectively.

Key Takeaways

Frequently Asked Questions

What is the biggest legal mistake companies make when deploying AI?

The most common mistake is inadequate contract terms, particularly regarding data ownership and usage rights, which can lead to substantial legal and operational risks.

Why is explainable AI important in business contracts?

Explainable AI allows businesses to trace decision-making processes, ensuring they can identify where a system went wrong, thus maintaining accountability and compliance with regulations.

How should businesses approach AI exit strategies?

Businesses should plan for data transition and retention, ensuring that systems can be decoupled without accumulating technical debt, maintaining continuity after ending AI vendor relationships.

How does AI affect job roles according to Rory O'Keeffe?

AI will not replace human jobs like legal professions but will change them. Those who skillfully integrate AI into their workflows will replace those who do not adapt.

What should founders consider before signing AI contracts?

Founders should ensure their AI contracts clearly define data use, ownership, anonymization processes, and have strategies for auditability, explainability, and planned exits to align with business risks and objectives.