Skip to main content
BlogOrganisational Change & Business Readiness

Six patterns from CEO conversations about AI

Digital Transformation Practice4 min read

The vignettes below are illustrative composites, not transcripts. I have stitched together patterns I keep hearing in advisory conversations with CEOs and board chairs across our markets, then anchored each one to the published evidence. Identifying details have been blurred or invented because the patterns matter and the specific people do not. What follows is six patterns, each drawn from many conversations, each backed by survey data I have actually read.

A note on framing before the patterns. The composites below describe directional changes only. Where I write that a bank "approved noticeably more loans with the same default rate" or that a legal team's throughput "rose meaningfully," I am describing the shape of an outcome, not a specific number from a specific client. Aggregate numbers in this piece come from cited sources (McKinsey, BCG, Deloitte, IBM, EY, KPMG, PwC, Gartner, Stanford, Slack, and others), not from confidential engagement work.

Pattern one. The CEO of a tier-two bank told me she was not worried about AI replacing her tellers. She was worried about her loan officers. The bank had built a credit scoring model and was approving noticeably more loans at a stable default rate. The team that maintained the model was small. The team that used to do the equivalent decisioning manually had been many times that. She had not laid anyone off. She had simply not backfilled when people left. McKinsey's 2024 State of AI survey reported that 41% of organisations using GenAI in HR expected workforce reductions in that function within three years, but for most other functions reductions trailed reallocation [1]. BCG's 2024 multi-country GenAI survey found 26% of executives describing AI as a threat to jobs, while only 15% of frontline staff said the same [2]. IBM's 2024 CEO Study found 47% of CEOs planning to hire for AI-related roles even while reporting overall productivity gains [3]. The non-backfill pattern is widely reported. Public layoff announcements understate what is actually happening to head counts.

Pattern two. The MD of a supermarket group asked how to think about AI in a business with thin margins. The honest answer was that most vendors pitching him would not deliver value at his cost base. The thing he could afford and that would actually move the business was inventory and shrinkage analytics. Not glamorous. Deloitte's 2024 State of Generative AI in the Enterprise survey found 67% of organisations citing scalability and 65% citing risk and governance as top adoption barriers, with only 27% reporting they had captured significant value at scale [4]. Gartner's 2024 GenAI hype cycle moved several enterprise application categories deeper into the trough of disillusionment, citing weak alignment between vendor capability and business problem [5]. Vendor pitches missing the highest-value problem is the modal outcome, not the exception.

Pattern three. The CEO of a regulated utility wanted to know whether AI could help with leak detection. The technical answer is yes, with caveats. The political answer was more interesting. He was about to commit a multi-year programme and the board was anxious. We spent more time on phasing than on modelling. UK water sector regulator Ofwat's 2024 PR24 final determinations expressed expectations for cautious, evidence-led innovation rather than headline transformation, and the Financial Times has tracked regulator scrutiny of utility capital programmes through 2024 and into 2025 [6]. Gartner's 2024 critical capabilities work on AI in regulated sectors found governance and audit explainability outranking model accuracy as procurement criteria [7]. Phasing matters more than methodology when the cost of a visible failure is set by the regulator, not the market.

Pattern four. The chair of an insurer asked, almost apologetically, whether AI was actually going to be a big deal or whether it was overhyped. She had been doing the job for many years and seen multiple technology waves not deliver on their promise. I told her my honest view: a real shift, larger than mobile, smaller than the public narrative claims, and concentrated in places that will surprise her. She asked which places. I told her the boring ones first. Underwriting. Claims. Internal document search. IBM's 2024 CEO Study found 64% of CEOs believing competitive advantage would depend on advanced GenAI, while only 38% felt their teams could turn that into operational gains [3]. KPMG's 2024 CEO Outlook reported AI as the top investment priority for 64% of CEOs, with 80% saying ROI would take three to five years to materialise [8]. Long-tenured leaders calibrating against prior waves end up closer to the data than executives who came up during the hype cycle. The relief I read in this conversation was the relief of a leader being told the boring places mattered most.

Pattern five. The CEO of a fintech wanted help thinking through what regulatory exposure his company would have if local regulators introduced AI-specific rules in the next eighteen months. The answer involved a careful read of the existing regulatory direction of travel and a candid acknowledgement that we did not know with confidence. He asked what he should do operationally. We worked through three scenarios and identified investment that would be defensible in any of them. He committed to that, and held back the additional investment that depended on a more permissive regulatory outcome. EY's 2024 CEO Outlook Pulse found 70% of CEOs identifying regulatory uncertainty as a top obstacle to AI investment, and 65% saying they were sequencing capital deployment around regulatory scenarios [9]. The Stanford AI Index 2024 documented a sharp rise in AI-related regulatory activity globally, with 25 new federal AI regulations in the United States alone in 2023 [10]. Investing what is defensible under multiple regulatory scenarios is the discipline. It does not get celebrated. It is exactly the right call.

Pattern six. The CEO of a partnership-style firm asked how to talk to partners about AI without losing them. About a third of the partnership was anxious about AI replacing aspects of work tied to professional identity. About a third was excited and pushing for aggressive deployment. About a third was indifferent and wanted clarity. He had been treating it as one conversation. We helped him separate it into three, with three different framings, in three different forums. PwC's 2024 Annual Global CEO Survey found 45% of CEOs doubting their company's viability over a ten-year horizon without reinvention, while only 27% had taken material steps in workforce planning [11]. Slack's 2024 Workforce Index found AI excitement falling and AI anxiety rising among knowledge workers between January and August 2024, with the share of workers expressing optimism dropping from 47% to 36% [12]. Partnership-style firms need separated AI forums because the population is bimodal at the top and the all-hands conversation flattens the signal.

What I take from these patterns together is that the most thoughtful CEOs are not asking whether AI matters. They have moved past that. They are asking how to invest without making the kind of mistake that is hard to reverse. PwC's 2024 CEO survey, IBM's CEO Study and KPMG's CEO Outlook all converge on that framing [3, 8, 11]. The boards still in the hype-or-doom register are mostly behind, even if they do not realise it yet.

Coderex advises CEOs and boards on the operational shape of these decisions: how to size investment under regulatory uncertainty, how to phase deployments where the cost of failure is set by a regulator, and how to design the AI conversation inside a partnership-style firm without losing the senior population at either end of the bimodal distribution.

Expect the 2025 CEO survey cycle from PwC, IBM, KPMG and McKinsey to show the calibration gap (the share of CEOs who believe AI matters versus the share who feel their teams can turn that belief into operational gains) widening rather than narrowing. Expect the non-backfill pattern to start surfacing in published headcount data over the next 12 to 24 months, particularly in financial services and professional services. Expect at least one regulator in a major jurisdiction to publish AI-specific phasing guidance for regulated sectors before 2027.

Methodology note: This article uses illustrative composite vignettes drawn from advisory conversations across multiple sectors and geographies. No vignette describes a specific client. All quantitative claims come from the cited sources, mostly published 2023 to 2024, including McKinsey, BCG, Deloitte, IBM, EY, KPMG, PwC, Gartner, Stanford HAI, and Slack. The patterns are mine. The numbers are theirs. Where the data is contested, I have flagged it.

References

13 sources, all verified at the time of writing

  1. [1]McKinsey & Company, 2024. The state of AI in early 2024: Gen AI adoption spikes and starts to generate value. McKinsey Quarterly. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai.
  2. [2]Boston Consulting Group, 2024. AI at Work 2024: Friend and Foe. BCG. https://www.bcg.com/publications/2024/ai-at-work-friend-and-foe.
  3. [3]IBM Institute for Business Value, 2024. 6 hard truths CEOs must face: 2024 CEO Study. IBM. https://www.ibm.com/thought-leadership/institute-business-value/c-suite-study/ceo.
  4. [4]Deloitte, 2024. State of Generative AI in the Enterprise: Q3 2024 report. Deloitte AI Institute. https://www2.deloitte.com/us/en/pages/consulting/articles/state-of-generative-ai-in-enterprise.html.
  5. [5]Gartner, 2024. Hype Cycle for Generative AI, 2024. Gartner Research. https://www.gartner.com/en/articles/hype-cycle-for-generative-ai.
  6. [6]Ofwat, 2024. PR24 final determinations: Overview. UK water sector regulator. https://www.ofwat.gov.uk/regulated-companies/price-review/2024-price-review/final-determinations/.
  7. [7]Gartner, 2024. Critical Capabilities for AI Trust, Risk and Security Management. Gartner Research. https://www.gartner.com/en/articles/what-is-ai-trism.
  8. [8]KPMG International, 2024. KPMG 2024 CEO Outlook. KPMG. https://kpmg.com/xx/en/our-insights/value-creation/kpmg-global-ceo-outlook-survey-2024.html.
  9. [9]EY, 2024. EY CEO Outlook Pulse: October 2024 edition. Ernst & Young. https://www.ey.com/content/dam/ey-unified-site/ey-com/en-gl/campaigns/ceo/documents/ey-gl-ceo-outlook-pulse-survey-09-2024.pdf.
  10. [10]Stanford Institute for Human-Centered Artificial Intelligence, 2024. Artificial Intelligence Index Report 2024. Stanford HAI. https://aiindex.stanford.edu/report/.
  11. [11]PwC, 2024. 27th Annual Global CEO Survey. PricewaterhouseCoopers. https://www.pwc.com/gx/en/issues/c-suite-insights/ceo-survey.html.
  12. [12]Slack and Workforce Lab, 2024. Workforce Index: September 2024 wave. Slack. https://slack.com/blog/news/the-fall-2024-workforce-index-shows-executives-and-employees-investing-in-ai-but-uncertainty-holding-back-adoption.
  13. [13]World Economic Forum, 2023. Future of Jobs Report 2023. WEF. https://www.weforum.org/publications/the-future-of-jobs-report-2023/.