Skip to main content
BlogOrganisational Change & Business Readiness

What CEOs Keep Asking Me About AI (And What I Actually Tell Them)

Digital Transformation Practice3 min read

Across the conversations I have had with senior leaders this year, six questions come up over and over. They are usually asked sideways, at the end of a meeting, after the slides are closed. I want to write down the questions and the answers I give. Not the boardroom-friendly ones. Most of what follows is supported by published surveys. Where I am leaning on practitioner judgement, I say so directly.

Question one. Are we behind? Almost always no. The companies asking me this are usually a year or so into pilots and convinced their competitors are running working systems. They mostly are not. McKinsey's 2025 State of AI puts overall AI use at 78% of organisations, but the share extracting meaningful financial value is much smaller [1][6]. BCG's 2024 work on the AI value gap put it bluntly: only about a quarter of companies report tangible value at any scale [2]. IBM's 2024 CEO study found 64% of chief executives say they are pushing GenAI faster than parts of their workforce can absorb [3]. Speed of announcement is not the same as speed of value.

Question two. Should we build or buy? Buy first. Build only the things that are actually strategic. Most companies do not need to train models. They need to use models well, and that means investing in data, evaluation pipelines, and integration. Gartner's 2024 forecast that around 30% of GenAI projects would be abandoned after proof of concept is sobering, and the build-cost ranges they reported, often six to seven figures for custom builds, match what we see in practice [8]. Deloitte's Q4 2024 enterprise survey found 68% of organisations had moved 30% or fewer of their GenAI experiments into production, with scaling and value measurement at the top of the barrier list [5].

Question three. How do we measure ROI? This is the question I am worst at answering, because the honest answer is that most early-stage AI deployments do not generate clean, attributable ROI in the first nine to twelve months. They generate option value. Brynjolfsson, Li, and Raymond's NBER study of more than five thousand customer-service agents found a 14% average productivity gain, with most of the benefit going to less-experienced workers [9]. McKinsey's 2024 update reports that the most common cost reductions show up in supply chain, service operations, and software engineering, but at activity level rather than as headline P&L wins [14]. KPMG's 2024 CEO Outlook found 64% of CEOs would invest in AI even without immediate returns, which I read as patience that has not yet been tested by a downturn [4].

Question four. What about the ethics? I have started asking this question back. Specifically what are you worried about? Bias in customer-facing decisions is one set of problems. Hallucination in legal or clinical text is another. Training data provenance and IP exposure are a third. Generic ethics frameworks rarely help anyone, in my view. The NIST AI Risk Management Framework gives a usable starting structure [12]. The EU AI Act adds real obligations on a staggered timeline, with prohibitions live since February 2025 and most high-risk requirements applying from August 2026 [13]. Map your specific systems to specific risks, then write the policy.

Question five. Will this replace my managers? Almost no. Will it change what they spend their time on? Almost certainly. McKinsey estimates GenAI could automate activities accounting for 60% to 70% of employees' time, with an annual productivity opportunity of $2.6 to $4.4 trillion [10]. The World Economic Forum's Future of Jobs 2025 projects 22% of jobs will face disruption by 2030, with 170 million roles created and 92 million displaced [11]. BCG's AI at Work 2024 survey of more than thirteen thousand workers found regular GenAI users report meaningful weekly time savings on routine tasks [15]. Based on what I see in client work, the mid-level manager job is half coordination and half judgement. AI is good at the first half. It is not good at the second.

Question six. When can we tell our shareholders we have an AI strategy? When you actually have one. A list of pilots is not a strategy. A real strategy connects pilots to a capability roadmap and answers where you intend to compete differently in eighteen months. If you can write that down, you have something to announce. If you cannot, the press release will not save you.

I do not know which of these questions will look obvious in two years and which will look hopelessly underbaked. Probably some of both. The companies that handle this well, in my experience, stop asking how to look credible on AI and start asking what they want to be capable of.

Coderex advises CEOs and boards on the operational shape of those decisions: where the firm wants to compete differently in eighteen months, the build-versus-buy line that fits their actual data and IP position, the three-bucket budget that separates capability from process automation from strategic bets, and the AI risk policy that treats a chatbot and an underwriting model as different problems.

Expect the 2026 CEO survey cycle from McKinsey, BCG, KPMG, and IBM to start showing the value-capture share rising past the current quarter as the rewiring work compounds, even while the lateness perception persists. Expect the EU AI Act's August 2026 high-risk obligations to surface a wave of governance retrofits across multinational portfolios. Expect at least one large public company to attribute a measurable margin shift to AI on an earnings call by mid-2027, which will reshape what the patient-but-untested CEO investment posture looks like.

Methodology note: This article draws on six recurring CEO and board questions I encounter in client work, then maps each to the most credible recent surveys and frameworks I could verify. Sources include McKinsey's State of AI work, BCG, KPMG, IBM, Deloitte, Gartner, the Stanford AI Index 2025, the WEF Future of Jobs 2025, the NIST AI RMF 1.0, the EU AI Act, and Brynjolfsson et al. (NBER 2023). Where I rely on practitioner observation rather than published data, I flag the sentence with phrases like "in my view" or "based on what I see in client work". Vendor surveys can carry methodological bias and I have noted that in the deep-dive version.

References

15 sources, all verified at the time of writing

  1. [1]McKinsey & Company, 2025. The state of AI: How organizations are rewiring to capture value. McKinsey Global Survey on AI. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai.
  2. [2]Boston Consulting Group, 2024. AI Adoption in 2024: 74% of Companies Struggle to Achieve and Scale Value. BCG Press Release. https://www.bcg.com/press/24october2024-ai-adoption-in-2024-74-of-companies-struggle-to-achieve-and-scale-value.
  3. [3]IBM Institute for Business Value, 2024. CEO decision-making in the age of AI: Six hard truths CEOs must confront. IBM. https://www.ibm.com/thought-leadership/institute-business-value/en-us/c-suite-study/ceo.
  4. [4]KPMG International, 2024. KPMG 2024 CEO Outlook. KPMG. https://kpmg.com/xx/en/our-insights/c-suite/ceo-outlook-survey-2024.html.
  5. [5]Deloitte, 2024. State of Generative AI in the Enterprise: Q4 2024. Deloitte AI Institute. https://www2.deloitte.com/us/en/pages/consulting/articles/state-of-generative-ai-in-enterprise.html.
  6. [6]Stanford Institute for Human-Centered AI, 2025. Artificial Intelligence Index Report 2025, Chapter 4: Economy. Stanford HAI. https://hai.stanford.edu/ai-index/2025-ai-index-report.
  7. [7]Gartner, 2024. Gartner Survey Finds Generative AI Is Now the Most Frequently Deployed AI Solution in Organizations. Gartner Press Release. https://www.gartner.com/en/newsroom/press-releases/2024-10-22-gartner-survey-finds-generative-ai-is-now-the-most-frequently-deployed-ai-solution-in-organizations.
  8. [8]Gartner, 2024. Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025. Gartner Press Release. https://www.gartner.com/en/newsroom/press-releases/2024-07-29-gartner-predicts-30-percent-of-generative-ai-projects-will-be-abandoned-after-proof-of-concept-by-end-of-2025.
  9. [9]Brynjolfsson, Erik et al., 2023. Generative AI at Work. NBER Working Paper No. 31161. https://www.nber.org/papers/w31161.
  10. [10]McKinsey & Company, 2023. The economic potential of generative AI: The next productivity frontier. McKinsey Global Institute. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier.
  11. [11]World Economic Forum, 2025. Future of Jobs Report 2025. World Economic Forum. https://www.weforum.org/publications/the-future-of-jobs-report-2025/.
  12. [12]NIST, 2023. Artificial Intelligence Risk Management Framework (AI RMF 1.0). U.S. National Institute of Standards and Technology. https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf.
  13. [13]European Parliament and Council, 2024. Regulation (EU) 2024/1689 (Artificial Intelligence Act). Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1689.
  14. [14]McKinsey & Company, 2024. The state of AI in early 2024: Gen AI adoption spikes and starts to generate value. McKinsey Global Survey on AI. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-early-2024-gen-ai-adoption-spikes-and-starts-to-generate-value.
  15. [15]BCG and Henderson Institute, 2024. AI at Work 2024: Friend and Foe. BCG. https://www.bcg.com/publications/2024/ai-at-work-friend-and-foe.