Category: Uncategorised

A supportive extension to the Next Digital Workforce blog series

Prepared for a C-suite audience

Executive thesis

The next digital workforce is no longer a metaphor for software adoption. It is becoming an operating reality in which human colleagues, copilots, autonomous agents, workflow orchestration tools and governed data platforms increasingly work side by side. For the C-suite, the implication is clear: the future workforce cannot be built only by reskilling today’s employees or buying agentic AI platforms. It must also be grown deliberately by developing a new generation of workers who know how to think, decide, collaborate and govern in an AI-rich workplace.

This is not simply a social-responsibility argument. It is a survival argument. Stanford’s 2025 AI Index reports that 78% of organisations were already using AI in 2024, up from 55% the year before, while private generative AI investment continued to grow strongly [1]. McKinsey’s 2025 workplace research finds that nearly all companies are investing in AI, but only 1% describe themselves as mature in deployment; the barrier is not employee readiness alone, but leadership’s ability to rewire the business around AI-enabled work [2]. Companies that fail to build the young talent pipeline for this new form of work risk creating elegant AI strategies without enough people capable of supervising, challenging, improving and responsibly scaling them.

What young people must practise and conquer

1. AI fluency, not superficial prompt literacy

Young people need to move beyond treating AI as a faster search engine or essay generator. They must understand the practical grammar of AI systems: how models are trained, where hallucination appears, how context windows, retrieval, tools and agents differ, and why outputs must be tested against evidence. Prompting remains useful, but the stronger capability is AI task design: defining the job to be done, selecting the right tool, sequencing human and machine steps, and knowing when not to automate.

2. Critical judgement at the jagged frontier

The most dangerous future employee is not the one who cannot use AI. It is the one who trusts it too much. Dell’Acqua and colleagues, working with Boston Consulting Group, showed that AI can improve performance on tasks inside its capability frontier but worsen outcomes on tasks just outside it. In their experiment, consultants using GPT-4 completed 12.2% more tasks and 25.1% faster on certain knowledge tasks, yet were 19% less likely to reach correct answers on a complex task outside the frontier [3]. Young workers therefore need disciplined scepticism: source checking, assumption testing, adversarial questioning, statistical intuition and the confidence to say, “the machine may be fluent, but it is not yet right.”

3. Agent management and digital delegation

Agentic AI changes the shape of junior work. Instead of only doing tasks, early-career employees will increasingly brief, monitor and coordinate task-performing systems. This demands a new craft: decomposing work into sub-tasks, setting success criteria, designing guardrails, auditing intermediate outputs, managing exceptions and escalating risk. Microsoft’s 2025 Work Trend Index describes the emergence of “Frontier Firms” where digital labour is embedded into strategy and workflows, and where leaders are actively considering both workforce upskilling and expanded capacity through digital labour [4]. Young people should therefore practise being “agent supervisors” before they become line managers.

4. Data, evidence and workflow literacy

Agentic AI is only as useful as the workflows and data into which it is embedded. Young people need practical literacy in data quality, provenance, bias, privacy, permissions, process mapping and measurement. They do not all need to become data scientists, but they do need to know why poor metadata, weak controls, broken integrations and unowned datasets turn promising AI into operational risk. OECD analysis makes a related point: most AI-exposed workers will not need specialist machine-learning skills, but AI will change the tasks they perform and increase the importance of management, business, digital, cognitive and emotional skills in exposed occupations [5].

5. Human advantage: communication, empathy and organisational intelligence

As AI absorbs more routine information processing, the premium on human interaction rises. Stanford’s Future of Work with AI Agents project finds that many tasks are likely to require human-agent collaboration, and that workers often prefer higher levels of human agency than technologists assume [6]. This matters. The young worker who can translate between engineers, customers, regulators, finance, operations and frontline staff will be more valuable than the young worker who merely produces technically plausible outputs. Communication, negotiation, ethical sensitivity, active listening and stakeholder management become core productivity skills, not optional soft skills.

6. Learning agility and resilient self-direction

The World Economic Forum’s Future of Jobs Report 2025 identifies AI and big data, networks and cybersecurity, and technology literacy as among the fastest-growing skills, while also highlighting creative thinking, resilience, flexibility, curiosity and lifelong learning as rising in importance [7]. This is a profound signal to educators, parents and employers. The young person entering work today should not be trained for one fixed role; they should be trained to reconfigure their role repeatedly as AI changes the economics of tasks.

Why companies must invest in young AI-native talent now

For many boards, the current workforce agenda has three pillars: enterprise AI strategy, technology investment and upskilling of existing employees. All three are necessary. None is sufficient. The missing fourth pillar is a deliberate young-talent strategy for the agentic era.

First, existing workforces carry the organisation’s tacit knowledge, customer memory and execution discipline. They must be upskilled, not discarded. Empirical evidence shows why this is valuable: Brynjolfsson, Li and Raymond found that a generative AI assistant in customer support increased productivity by 14% on average, with particularly large gains for novice and lower-skilled workers, suggesting that AI can transmit elements of expert practice and shorten the learning curve [8]. Properly used, AI can make early-career development faster and more equitable.

Second, however, incumbents alone cannot refresh the organisation’s mental model quickly enough. Young people who have grown up with AI-enabled learning, media, coding, creation and collaboration may bring a more natural sense of human-machine teaming. They can challenge legacy process assumptions, prototype agentic workflows, test new customer interfaces and expose where governance is too slow or too brittle. This does not mean putting youth in charge of enterprise risk. It means pairing young AI-native talent with experienced domain leaders in structured apprenticeship models.

Third, companies that ignore young talent will damage their future option value. Agentic AI will create new roles: agent operations manager, AI workflow designer, model-risk analyst, synthetic-data steward, AI product ethicist, human-agent experience designer, automation assurance lead and many more. These jobs will not be filled adequately by waiting for the market to produce mature candidates. They must be cultivated through internships, graduate rotations, apprenticeships, university partnerships, hackathons, internal academies and board-visible talent pathways.

The C-suite action agenda

The practical recommendation is to treat young AI-ready talent as a strategic resource class alongside cloud, data, cybersecurity and AI platforms. This requires five moves.

First, define an enterprise AI skills taxonomy that distinguishes AI fluency, domain expertise, workflow design, governance, data stewardship, cyber awareness, critical reasoning and human collaboration. Avoid vague language such as “digital native”; measure real capabilities.

Second, create agentic apprenticeships. Pair graduates and early-career employees with business owners to redesign real workflows using governed AI agents. Require every project to include a business metric, a risk assessment, a human-in-the-loop design and a lessons-learned artefact.

Third, build a “young talent plus expert mentor” operating model. The strongest teams will combine youthful experimentation with experienced judgement. This mirrors the emerging reality of the next digital workforce: humans at different career stages, agents and systems working in supervised partnership.

Fourth, invest in AI governance as an enabler, not a brake. Young workers must be taught responsible use from day one: privacy, explainability, bias, security, intellectual property, record keeping and escalation. Governance should be embedded into tools and workflows so that responsible innovation is easy to practise.

Fifth, make the CEO and CHRO jointly accountable. This agenda cannot sit only in learning and development or the CIO’s office. It affects productivity, risk, culture, succession, innovation and future competitiveness. The board should ask for quarterly reporting on AI skills coverage, young-talent pipeline, adoption quality, workflow redesign, risk incidents and measurable business outcomes.

Conclusion

The agentic AI era will not reward companies that simply buy tools, announce strategies and hope the workforce adapts. It will reward companies that engineer a new compact between people and machines. Upskilling the existing workforce protects today’s performance. Investing in young AI-native talent protects tomorrow’s relevance. Building both in parallel is how companies maximise their chances of survival, renewal and growth.

For the C-suite, the choice is stark. Either develop the next digital workforce deliberately, or inherit a future in which the organisation has powerful agents, ageing skills, weak supervision and no credible talent bridge between strategy and execution. The companies that win will be those that understand that the scarcest resource in the agentic AI era is not the model. It is the human capability to direct it wisely.

Selected references

[1] Stanford HAI, The 2025 AI Index Report, 2025. https://hai.stanford.edu/ai-index/2025-ai-index-report

[2] McKinsey & Company, Superagency in the workplace: Empowering people to unlock AI’s full potential, 2025. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work

[3] Dell’Acqua, F. et al., Navigating the Jagged Technological Frontier, Harvard Business School Working Paper No. 24-013, 2023. https://www.hbs.edu/faculty/Pages/item.aspx?num=64700

[4] Microsoft WorkLab, 2025: The year the Frontier Firm is born, Work Trend Index, 2025. https://www.microsoft.com/en-us/worklab/work-trend-index/2025-the-year-the-frontier-firm-is-born

[5] Green, A., Artificial intelligence and the changing demand for skills in the labour market, OECD Artificial Intelligence Papers No. 14, 2024. https://doi.org/10.1787/88684e36-en

[6] Stanford SALT Lab, Future of Work with AI Agents, 2025. https://futureofwork.saltlab.stanford.edu/

[7] World Economic Forum, The Future of Jobs Report 2025, 2025. https://www.weforum.org/publications/the-future-of-jobs-report-2025/

[8] Brynjolfsson, E., Li, D. and Raymond, L., Generative AI at Work, NBER Working Paper No. 31161, 2023. https://www.nber.org/papers/w31161

[9] Noy, S. and Zhang, W., Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence, Science, 2023. https://www.science.org/doi/10.1126/science.adh2586

Why AI transformation now belongs on the CEO agenda – and why the decisive advantage will be human, not merely technical

The argument

Over the past few years, the AI conversation in business has moved through three distinct phases. The first was fascination: generative AI as an extraordinary instrument for writing, searching, summarising and coding. The second was experimentation: pilots, sandboxes, copilots, innovation days and executive theatre. The third phase, the one now arriving, is more serious. AI is becoming a test of organisational speed. Not speed as haste, not speed as uncontrolled adoption, and certainly not speed as buying every fashionable tool in the market. I mean the deeper speed of an enterprise: the capacity to sense change, decide intelligently, redesign work, govern risk and learn faster than the environment is changing around it.

This matters particularly for large multinational enterprises carrying accumulated technical debt in tools, infrastructure, data and management practice. Many of these organisations were built for scale, resilience and control, not for continuous recomposition. Their ERP estates, data lakes, cloud migrations, procurement cycles, cyber controls, risk committees and legacy applications were not designed for a world in which intelligence is becoming embedded in every workflow. AI has exposed what was already true: the limiting factor is rarely the model. The limiting factor is the operating model.

This is the point I have tried to make in my recent writing on Generative AI and Agentic AI. The interesting question is no longer whether a model can produce an acceptable answer. The question is whether the organisation can turn that answer into a governed action, at scale, in context, with accountability. Agentic AI intensifies the issue because it shifts the discussion from tools that assist people to systems that initiate, plan, call other systems, execute tasks and learn from feedback. That is not a software upgrade. It is a challenge to the organisation’s metabolism.

What has changed

The empirical evidence now shows two truths moving together. First, adoption has accelerated dramatically. Stanford’s 2025 AI Index reported that 78% of organisations were using AI in 2024, up from 55% the previous year, while generative AI investment continued to expand significantly. McKinsey’s 2025 State of AI survey similarly describes wider use of AI and agentic AI, but also notes that many organisations are still struggling to move from pilots to scaled economic value. The pattern is clear: AI has crossed the adoption threshold, but not yet the transformation threshold.

Second, we are learning that value does not arrive evenly. Brynjolfsson, Li and Raymond’s research on generative AI in customer support found average productivity improvements of about 14%, with the largest gains accruing to less experienced workers. Dell’Acqua and colleagues, in their study with BCG consultants, described the ‘jagged technological frontier’: AI can lift performance significantly for tasks within its frontier and degrade performance for tasks outside it. This is crucial for boards. AI is not a universal accelerator. It is a conditional accelerator. It rewards judgement, task decomposition, good data, domain context and feedback. It punishes blind delegation.

This is why so many pilots disappoint. MIT’s 2025 GenAI Divide report argued that many enterprise initiatives fail because they are brittle, poorly integrated into daily work and unable to learn from context. Deloitte’s 2025 enterprise research similarly points to rising investment alongside elusive returns. IBM’s 2025 CEO research found that rapid investment has often created disconnected technology, while IBM’s 2026 CEO research reported that 83% of CEOs believe AI success depends more on people adoption than on technology itself. The message is no longer subtle: AI transformation fails when it is treated as deployment rather than rewiring.

From digital transformation to intelligent transformation

For thirty years, business transformation was largely about digitising existing processes. We put forms online, moved workloads to cloud, integrated channels, automated back offices and introduced analytics. Much of this was valuable, but it often preserved the inherited shape of the organisation. AI is different because it changes the unit of work. It can read, reason, generate, classify, converse, code and increasingly orchestrate. In agentic form, it can become a new participant in the enterprise operating system.

That creates a dangerous temptation: to insert AI into old processes and call it transformation. A CEO with a heavily indebted technology estate should resist this. If the process is broken, AI will accelerate the brokenness. If the data is fragmented, AI will make the fragmentation visible. If accountability is unclear, AI will amplify ambiguity. If middle management has been trained to protect functional boundaries, AI will not magically create cross-enterprise flow. The organisation will merely become faster at revealing its own incoherence.

The better question is: where are the enterprise constraints that AI now makes negotiable? Which approvals exist because information used to be scarce? Which reports exist because systems could not explain themselves? Which roles exist to reconcile data that should never have been inconsistent? Which customer journeys are slow because the organisation is divided by internal functions rather than external outcomes? Which technical debt has been tolerated because the cost of change was historically too high? AI changes the cost curve of coordination, but only if leadership is willing to challenge the contracts embedded in the organisation.

What it means to build organisational speed

Organisational speed is not the same as moving quickly. Many organisations are already fast in the wrong places. They can launch pilots quickly, buy tools quickly and issue press releases quickly. The more valuable form of speed has five characteristics.

The first is speed of sense-making. Leaders need the ability to detect where AI is changing customer expectations, cost structures, risk profiles and competitive boundaries. This requires external scanning, internal telemetry and board-level fluency. A board that treats AI as a technology topic will be late; a board that treats AI as a strategic discontinuity has a chance.

The second is speed of decision. AI opportunities decay when they are trapped in committees designed for yesterday’s risk. This does not mean weakening governance. It means designing governance that is proportionate, informed and close to the work. Responsible AI, security, data protection and model assurance must be built into the delivery system, not bolted on as a final inspection.

The third is speed of learning. Organisations must move from pilot culture to learning culture. A pilot asks whether a tool works. A learning system asks what changed in the work, what was adopted by people, what risk emerged, what data improved, what should be stopped and what should scale. This is where many enterprises are weakest. They accumulate experiments without compounding knowledge.

The fourth is speed of integration. The next advantage will not come from isolated copilots. It will come from connecting models to workflows, data, controls, APIs, human review, cyber policy, auditability and business outcomes. This is where technical debt becomes strategic debt. Legacy infrastructure is not merely an IT inconvenience; it is a brake on organisational learning.

The fifth is speed of trust. People will not adopt systems they do not understand, cannot challenge or believe are being used against them. Trust is not soft. It is the lubricant of transformation. Without it, employees route around new tools, managers preserve old behaviours and the organisation creates a theatre of adoption while real work continues elsewhere.

Why the deepest transformations are about people

BCG has often framed AI value through a 10-20-70 logic: a smaller proportion lies in algorithms, more in technology and data, and the majority in people, process and change. Whether one accepts the exact numbers or not, the principle is right. The transformation is ultimately human because work is social before it is technical. Decisions are made by people, exceptions are handled by people, customers trust people, risk is owned by people and culture is transmitted by people.

The World Economic Forum’s Future of Jobs Report 2025 expects 39% of workers’ core skills to change by 2030 and estimates that, in a workforce of 100 people, 59 will need training before the end of the decade. That is not an HR footnote. It is a balance sheet issue. Skills are now a strategic asset class. The enterprise that cannot reskill quickly cannot transform quickly. The enterprise that cannot redesign roles cannot capture AI value. The enterprise that treats people as recipients of change rather than authors of change will lose the very intelligence it needs.

This is particularly true for middle management. In many large enterprises, middle managers are the translation layer between strategy and work. They can either become the accelerators of AI transformation or its immune system. If they are excluded, threatened or left untrained, they will slow the transformation in rational self-defence. If they are equipped to redesign work, coach teams, manage risk and interpret AI outputs, they become the most important agents of speed.

The same is true for frontline expertise. AI systems require context. They need to learn from the people who know where processes fail, where customers become frustrated, where data is misleading, where policies contradict reality and where exceptions actually occur. In this sense, AI does not remove the need for human intelligence; it increases the premium on human judgement. The future enterprise is not a machine with people attached. It is a human institution with new cognitive infrastructure.

The CEO agenda

For the CEO of a large multinational enterprise, the practical implications are stark. First, do not allow AI to become another layer of technical debt. Every AI investment should be tested against architecture, data lineage, cyber posture, model governance and integration into real work. Second, move from use-case enthusiasm to capability building. The question is not how many pilots are running, but whether the organisation is building reusable data products, model assurance, workflow orchestration, talent pathways and decision rights. Third, make adoption a leadership discipline. Usage statistics are not enough; measure changes in cycle time, quality, customer outcomes, employee confidence and risk controls.

Fourth, create a strategic map of what must be rewired. Some processes should be automated, some augmented, some eliminated and some protected because human judgement is the source of value. Fifth, put people at the centre without romanticising the status quo. People-led transformation does not mean avoiding difficult choices. It means making those choices with clarity, fairness, participation and investment in capability.

My own view is that AI transformation is entering its second act. The first act was about possibility. The second is about organisational character. The winners will not be the firms with the most pilots, the largest tool catalogue or the loudest AI narrative. They will be the firms that build speed without losing judgement, automate without abandoning accountability, and use AI to enlarge human agency rather than merely reduce human cost.

That is why this is a CEO issue. Technical debt, process debt and skills debt have converged. AI has made the hidden friction of the enterprise visible. The question for the board is not whether the organisation should adopt AI. That decision has already been made by the market. The question is whether the organisation can rewire itself quickly enough, wisely enough and humanely enough to turn intelligence into advantage.

Selected references

Brynjolfsson, E., Li, D. and Raymond, L. R. (2023/2025), ‘Generative AI at Work’, NBER Working Paper 31161 and Quarterly Journal of Economics.

Dell’Acqua, F. et al. (2023/2025), ‘Navigating the Jagged Technological Frontier’, Harvard Business School / Organization Science.

Deloitte (2025), The State of Generative AI in the Enterprise.

IBM Institute for Business Value (2025), CEO Study: CEOs Double Down on AI While Navigating Enterprise Hurdles.

IBM Institute for Business Value (2026), CEO Study: CEOs are Reshaping C-suite Roles for the AI Era.

McKinsey & Company (2025), The State of AI: Global Survey 2025.

MIT NANDA (2025), The GenAI Divide: State of AI in Business 2025.

Stanford HAI (2025), Artificial Intelligence Index Report 2025.

World Economic Forum (2025), The Future of Jobs Report 2025.

Source URLs consulted: 

https://hai.stanford.edu/ai-index/2025-ai-index-report

https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

https://www.nber.org/papers/w31161

https://www.hbs.edu/faculty/Pages/item.aspx?num=64700

https://www.deloitte.com/uk/en/issues/generative-ai/state-of-generative-ai-in-enterprise.html

https://newsroom.ibm.com/2025-05-06-ibm-study-ceos-double-down-on-ai-while-navigating-enterprise-hurdles

https://newsroom.ibm.com/2026-05-04-ibm-study-ceos-are-reshaping-c-suite-roles-for-the-AI-era

https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf

https://www.weforum.org/publications/the-future-of-jobs-report-2025

Why boards and CEOs must look beyond application-layer productivity and ask which hidden contracts in the business technology stack are about to be renegotiated by Agentic AI

The wrong conversation is winning

In recent posts I have argued that Generative AI is not simply a tool for producing more words, more code, more images or more slideware. I have also argued that Agentic AI is not merely “automation with a better user interface”. Agentic AI changes where agency sits in the organisation. It moves decision, orchestration and execution into systems that can observe, reason, act, test, learn and recover. That is why I keep returning to governance, accountability and economic impact. The technology is interesting, but the transfer of agency is the point.

Most board conversations, however, are still stuck in the most visible part of the game. They ask whether developers will be 20% or 40% faster. They ask whether a copilot can clear the technology backlog. They ask whether the customer service team can answer more tickets with fewer people. These are legitimate questions, but they are scoreboard questions. They tell us whether today’s team is running harder under today’s rules. They do not tell us whether the rules of the sport are being rewritten.

The deeper AI story is not that software teams are being given better boots. It is that the pitch, the rules, the coaching staff, the refereeing system, the medical science, the broadcast model and the business model of the club are all becoming writable at the same time. Boards that only watch the striker will miss the change in the entire league.

From productivity story to paradigm story

For fifty years, the technology stack has behaved like a professional sport with rigid divisions of labour. The players play, coaches coach, referees referee, grounds teams maintain the surface, broadcasters package the spectacle, and owners negotiate commercial rights. Each role has its own contract, language, incentives and power base. In technology, those contracts are the application interface (API’s), the runtime, the operating system, the compiler, the instruction set and the silicon. We have treated them as fixed because changing them together was prohibitively expensive.

The application developer did not negotiate with the chip. The chip designer did not negotiate with the customer journey. The compiler expert did not sit in the boardroom discussing revenue leakage. The contracts were not laws of physics. They were historical settlements that lasted so long they started to look inevitable.

Large language models (LLM’s) and Agentic systems now challenge that assumption. They can work across the boundaries that humans built around expertise. They can write application code, generate compiler optimisations, synthesize kernel extensions, assist chip design, support formal verification and help redesign silicon layouts. They do not respect the old dressing-room politics because they did not grow up inside them. They ask a question most organisations have been structurally unable to ask: what happens if several layers of the game can be changed together?

That is why this is not primarily a productivity story. It is a paradigm story. Productivity improves the existing formation. Paradigm change alters what a formation is.

The evidence is already on the field

This is not science fiction. At the compiler layer, Meta’s LLM Compiler was trained on 546 billion tokens of LLVM intermediate representation and assembly code. Its published results show 77% of the optimisation potential of an autotuning search and 45% disassembly round-trip accuracy from x86 and ARM assembly back into LLVM IR. In plain English, models are learning the language beneath the language most software teams discuss.

At the operating-system layer, Kgent, presented at the ACM SIGCOMM 2024 eBPF workshop, translates natural language prompts into eBPF programs for the Linux kernel. Related 2025 work on agentic operating systems uses LLM agents to analyse workloads, synthesize eBPF scheduling policies and deploy them through sched_ext, reporting up to 1.79x performance improvement and a 13x cost reduction compared with naive agent approaches. The “grounds team” of the computing world is no longer only repairing turf; parts of the turf-management strategy are becoming agentic.

At the chip-design layer, NVIDIA’s ChipNeMo explores domain-adapted large language models for industrial chip design, including engineering assistant use cases, EDA script generation and bug summarisation. Google DeepMind’s AlphaChip uses reinforcement learning to generate chip floorplans in hours rather than the weeks or months traditionally required, and Google says these layouts have been used in multiple generations of TPUs and in Axion CPU work. At the instruction-set layer, RISC-V research on the XiangShan Nanhu processor extends the instruction set for LLM vector dot-product acceleration, reporting more than four times scalar-method speed on that core operation.

The important point is not that any one of these examples wins the match alone. The point is that they are appearing at multiple layers at once. This is the early sign of what I have called consurgence: new properties rising together because the conditions have changed together.

A sporting metaphor for the boardroom

Imagine a football club that has spent decades improving marginally within known constraints. It hires better strikers, buys improved boots, upgrades nutrition, deploys data analysts and improves ticketing. Each initiative is useful. Some are very profitable. But all assume the same pitch, the same league rules, the same broadcast economics and the same match-day model.

Now imagine a new class of agentic system arrives that can redesign the training plan, simulate match tactics, change the playing surface, alter recovery science, generate new scouting models, negotiate media packages, optimise stadium logistics and propose rule changes to the league. The club that asks only, “Can our striker score 10% more goals?” has misunderstood the opportunity. The better question is, “Which assumptions about the game have become negotiable?”

That is where companies are today. The visible AI market is crowded with applications: copilots, assistants, chatbots, content tools, workflow wrappers and SaaS add-ons. This is the striker market. It is exciting, noisy and overfunded. But the deeper advantage will come from the less glamorous parts of the club: the academy, sports science, data infrastructure, playing surface, referee technology, transfer analytics and league governance. In technology terms, those are compilers, runtimes, verification, operating systems, open silicon tooling, capability security and the governance of agentic action.

Why this matters to CEOs and boards

Boards do not need to become chip designers. CEOs do not need to run compiler teams. But they do need to understand where durable advantage is likely to form. If AI simply improves the application layer, then the winning strategy is adoption speed: choose tools, train staff, measure productivity and manage risk. If AI rewrites the layers beneath the application, the winning strategy is different: identify which hidden contracts shape your economics and decide whether to defend, renegotiate or escape them.

For a telco, the hidden contract may be between network operations, customer experience and vendor equipment roadmaps. For a bank, it may be between risk models, legacy core systems, regulatory reporting and the human sign-off process. For a logistics company, it may be between routing, fleet maintenance, insurance, carbon reporting and customer promises. For a public-sector body, it may be between policy intent, operational data, procurement rules and citizen outcomes. Agentic AI is not powerful because it “does tasks”. It is powerful because it can coordinate across these boundaries and reveal which boundaries were artificial.

This is also why AI governance cannot be reduced to ethics theatre or model-risk paperwork. My position has consistently been that governance must be an operating discipline. When agents can act across systems, governance must define permission, accountability, evidence, reversibility and escalation. The question is not only “Was the model fair?” It is also, “Who authorised the agent to change the play, under what constraints, with what audit trail, and who can stop it when the match changes?”

The capital allocation mistake

The current investment pattern is heavily weighted towards the obvious. Most corporate pilots and much venture capital crowd around application-layer tools because they are easy to demonstrate and easy to sell. The board can see the demo. The CFO can imagine the headcount saving. The press release almost writes itself.

But if the real renegotiation is happening below the application layer, then the smarter capital question changes. Companies should still invest in adoption, but not confuse adoption with advantage. The layers that deserve more attention are the systems that translate new compute into usable business capability: compiler and runtime infrastructure, formal verification, secure agent execution, open silicon design tooling, heterogeneous compute operating models, and hardware-enforced capability boundaries such as CHERI and Arm Morello. These are not fashionable board topics, but neither were cloud control planes, mobile app stores or semiconductor supply chains until they determined who captured the margin.

The UK policy picture illustrates the risk. The AI Opportunities Action Plan and subsequent updates rightly emphasise compute, AI Growth Zones, public-sector adoption, data assets, sovereign capability and skills. UKRI’s 2026 AI strategy commits more than £1.6 billion of targeted AI funding, while the Sovereign AI Unit is backed by up to £500 million. These are serious commitments. Yet the public framing still leans towards bedrock compute and visible adoption. The middle layers, where contracts become moats, receive far less public attention.

That is not a criticism of intent. It is a challenge to worldview. If you believe AI is another wave to be added to the existing stack, the current allocation is coherent. If you believe AI makes the stack itself writable, then the allocation looks incomplete.

The board agenda should change

The practical board agenda should now include five questions.

First: which technology contracts have we treated as fixed because changing them was historically too expensive? These may be vendor contracts, data architectures, operating models, platform dependencies, regulatory workflows or pricing models.

Second: where are we confusing productivity with strategic advantage? A 30% faster process inside a soon-to-be-obsolete model is not transformation. It is a better warm-up before the wrong match.

Third: which agents will be allowed to act, not merely advise? The moment an agent can trigger workflow, move money, reconfigure infrastructure, change a customer promise or alter a policy, the governance model must mature from advisory AI to accountable agency.

Fourth: are we investing only in the striker, or also in the academy, pitch, data room and rulebook? The visible applications matter, but they may become commoditised quickly. The less visible layers may determine lock-in, security, resilience and strategic optionality.

Fifth: what would we do differently if the cost of coordinated change fell by an order of magnitude? This is the question that forces leadership to think beyond process improvement and into new operating models.

A different game is forming

The danger for incumbents is not stupidity. It is success. Incumbents are optimised around the current contracts. Their management systems, budgets, partner ecosystems and board reporting all assume the sport remains recognisable. They will naturally describe change as incremental because their advantage is built inside the old rulebook.

History is not kind to those who manage the visible crisis and miss the structural one. Nokia, Kodak, Xerox and Blockbuster were not short of smart people. They were short of timely courage. They read the current scoreboard well and misread the future league.

The same pattern is visible now. Many leaders are debating whether AI can patch the roof of the current software business model: per-seat pricing, SaaS renewals, developer velocity, customer-service automation. Those discussions matter, but they are not enough. Agentic AI points towards a world where the unit of value may not be the application, the seat or even the workflow. It may be the agentic capability to orchestrate outcomes across systems, with governance strong enough to make that capability trusted.

The CEO takeaway

The old question was: how much faster can AI make our people? The better question is: which assumptions in our business have become negotiable?

The old question was: where can we add GenAI to the existing application estate? The better question is: where are agents about to coordinate across boundaries that our organisation chart still treats as sacred?

The old question was: should we invest in AI tools? The better question is: are we investing in the layers where advantage will compound, or only in the layer where the crowd has gathered?

In sport, the great clubs do not only buy stars. They build systems: academies, analytics, coaching, medical capability, recruitment intelligence, stadium economics and culture. They understand that sustainable advantage is rarely found only in the moment the ball hits the net. It is created in the conditions that make that moment repeatable.

That is the lesson for boards and CEOs. Generative AI has made the visible play more productive. Agentic AI is beginning to change who calls the play, how the play is executed, and whether the rules themselves can be rewritten. The scoreboard will not tell you that early enough. You have to look at the whole game.

My advice is simple: stop watching only the striker. The game is changing under the pitch.

Selected references and further reading

• Meta LLM Compiler: Foundation Models of Compiler Optimization

• NVIDIA Research: ChipNeMo – Domain-Adapted LLMs for Chip Design

• Google DeepMind: How AlphaChip transformed computer chip design

• Kgent: Kernel Extensions Large Language Model Agent, ACM SIGCOMM eBPF 2024

• Towards Agentic OS: An LLM Agent Framework for Linux Schedulers

• RISC-V XiangShan Nanhu LLM acceleration research

• UK Government: AI Opportunities Action Plan – One Year On

• UKRI Artificial Intelligence Research and Innovation Strategic Framework

• Arm Morello Program and CHERI capability technology