The management issue is not whether AI will change work, but who is allowed to decide which work changes first. For CIOs, the practical challenge is building a decision model that links business capability redesign, workforce implications and technology investment. Without that linkage, AI initiatives tend to fragment into isolated pilots, competing priorities and unclear ownership for outcomes.
That makes governance the real lever. CIOs need explicit decision rights with HR, finance and business leaders over three questions: which processes are candidates for automation or augmentation, which roles must be redesigned, and which controls are required before scaling. The trade-off is speed versus assurance. Fast experimentation can surface value quickly, but it also increases the risk of creating uneven policies, hidden dependencies and inconsistent employee experience across functions.
The portfolio implication is equally important. AI programmes should not be funded as a generic innovation bucket; they need to be ranked against near-term productivity, workforce readiness, risk exposure and process criticality. A useful next question for executive teams is: where will AI remove cost, where will it raise service quality, and where will it simply shift work to another part of the organisation? Those are different business cases and should not compete on the same metric.
Finally, adoption should be treated as an operating-model change, not a training event. The hardest work is often clarifying new accountability for decisions made with AI support, and defining what good human oversight looks like in practice. CIOs and HR leaders should ask: which roles need redesign now, which need reskilling, and which controls will prove that the organisation can use AI responsibly at scale?
-
How work is changing across roles and functions.
-
How systems must evolve to support AI-augmented work.
-
What skills workers will need to remain relevant.
Job disruption and workforce shifts
Last November, aA 2025 Stanford University study led by Dr. Erik Brynjolfssonย using ADP data from millions of ADP payroll records found that AI is already driving labor market shifts. Early-career workers in AI-exposed occupations are experiencing a 16% decline in employment, while employment for experienced workers remains stable so far . To be clear, employment changes are concentrated in occupations where AI automates rather than augments labor. Without question, AI will affect tasks, occupations, and industries in different ways, replacing work in some, augmenting others, and transforming still others. Professions already impacted affected at the hiring level include software developers and customer service representatives. More experienced workers have not been disrupted at the same rate, despite being less likely to embrace AI to augment their work. Generative AI tools like Claude and Open AI models are already demonstrating gains in personal productivity. The question that CIOs, along with their HR partners, need to answer is which business processes and tasks within the enterprise will be automated, augmented, or changed, and over time, what work looks like if agents handle the execution. In the longer-term agentic world, humans will be responsible for architecting work, putting together governance structures (guidelines, guardrails, and standards), and managing how agents do their execution. In this phase, Ian Beacraft said in his South by Southwest talk a few weeks ago, we move to agentic organizations.Which work gets automated, augmented, rebuilt
To navigate the workforce shift, CIOs should task their enterprise architects to take their maps of business capabilities and business processes and determine which will be automated, augmented, or changed. In practice, this makes enterprise architecture the mechanism for redesigning work. In many cases, this should be done using future-state maps that reflect how AI can transform operating models and create new value propositions. With these in hand, CIOs, along with their HR partners and AI-skilled architects, should evaluate job skills, determine which can be automated or augmented with AI, and align them to job descriptions.Role of enterprise architects
Enterprise architects can help by relating maps of business capabilities, skills, and job descriptions. To be clear, EAs are in an assist role; they should not own the workforce or job redesign directly. Instead, EAs should connect the dots across business capabilities, processes, systems, data, the operating model and governance, helping to inform role and skills changes alongside HR and the business. This collective effort should result in two things: first, identifying highly automated jobs; second, defining new job classifications that will combine job tasks from partially automated jobs or for roles that will manage agent-driven work and performance. This may be the most important job that enterprise architects ever perform — a rebuilding of the entire enterprise.Systems must support AI-augmented work
With this completed, the next logical question to consider is how systems should be designed to better support augmented jobs. For these positions, the question that CIOs, CHROs, EAs and CEOs need to consider is what systems must be able to do to support augmented work — and where they fall short today. These are big questions that must be answered collaboratively. Once again, enterprise architects need to take center stage.12 skills CIOs say workers need to stay relevant
Lastly, I asked CIOs about the skills that workers should develop to be relevant in an AI-driven future. Their answers were synthesized into 12 skill recommendations.-
AI fluency. Understand how AI models work — how they ingest, process and validate data — and where their limitations lie.
-
Human judgment. Apply critical thinking to assess AI outputs, especially when something feels off or incomplete.
-
Problem-solving. Ability to frame the right questions and use AI to accelerate better, more informed decisions.
-
Ethical responsibility and AI safety awareness. Understand how AI is used responsibly, with attention to bias, risk, accountability and governance.
-
Adaptability. Ability to continuously adjust to rapidly evolving tools, workflows and business expectations.
-
Continuous learning mindset. Commitment to ongoing skill development as AI reshapes roles and required capabilities.
-
Business acumen. Understand core business goals, processes and value drivers to ensure AI delivers meaningful outcomes.
-
Process and systems thinking. Ability to reimagine workflows end-to-end — moving from isolated tasks to integrated, AI-enabled outcomes.
-
Creativity and innovation. Identify new data sets, use cases and ways AI can unlock value — not just optimize existing work.
-
Communication and translation skills. Bridge technical and business worlds by explaining AI concepts in clear, actionable terms.
-
Cross-functional collaboration. Work effectively across IT, HR and business units as AI becomes embedded in every function.
-
Outcome orientation. Focus on building systems that deliver predictive insights and measurable business impact.
What CIOs can’t afford to get wrong
This article argues that this is a moment for CIOs to step up and partner deeply across the organization. It also highlights a critical opportunity for enterprise architects to help define the path forward. Delivering on this opportunity will require discipline and strong collaboration. Success will depend on building the right mix of skills. The winners in the AI era won’t be defined by technical depth alone, but by their ability to combine human capabilities — judgment, creativity, ethics — with AI as a partner to drive business outcomes.Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.