AI is forcing companies to redesign technology organisations from the inside out as agents take on more work. The shift places CIOs at the centre of enterprise change, alongside CEOs, business leaders and chief human resource officers. Technology leaders face pressure to reduce run costs, fund AI rollouts, manage geopolitical exposure, strengthen vendor returns and reshape workforces around human–agent teams. McKinsey research from April 2026 links stronger performance with deeper strategic involvement by technology leaders: two-thirds of top-performing companies have technology leaders deeply involved in enterprise strategy, compared with 52% of other organisations. The central challenge is how to hire selectively, build skills that AI cannot replace and reset vendor models to deliver measurable value.
Pressures Redefine Technology Leadership
Technology leaders face a tighter operating environment while demand for innovation continues to rise across functions. CIOs must reduce spending needed to keep technology infrastructure running while freeing budget for change initiatives such as AI deployments. That tension is growing because AI also requires robust infrastructure. Technology spending is therefore rising across many companies, with top-performing companies especially likely to increase budgets. A quarter of these companies plan technology budget increases of more than 10% this year, compared with 3% of other companies.
Geopolitical uncertainty adds further complexity. Global capability centres have moved beyond cost-focused delivery to become engines of innovation and product development. That evolution increases exposure to regulation, data security concerns and decisions on where talent should sit. Offshoring choices that once centred on cost now require a wider view of resilience and total cost of ownership. Vendor relationships and workforce planning add pressure from two more directions. AI requires buyers and technology vendors to rethink value propositions, technology stacks and operating models. Demand for AI talent remains intense, even as agentic AI performs more lower-level development work and increases the need to reskill technology employees.
Selective Hiring and Skills Shape Teams
Agentic AI changes the logic of technology hiring because routine development and maintenance tasks no longer justify the same staffing model. The central question shifts from expanding engineering capacity to placing human judgement where it matters most. Companies still need AI specialists, but they also need to reskill existing teams and reconsider which roles create the greatest value.
Leading organisations hire fewer technologists overall and apply greater selectivity to each appointment. Demand rises for senior engineers, architects, product managers and designers who can define standards and orchestrate work across internal teams, external teams, vendors and agents. The case for hiring large numbers of junior developers weakens as agentic AI takes on tasks previously assigned to these roles. With agents as tools, expert technologists can become several times more productive than less experienced peers, while the pay differential remains modest relative to the output gap.
Must Read: Clinical AI Scribes Face Speech Variability Test
A global healthcare company reorganised hundreds of technologists into a product-based operating model, accelerating delivery and creating clearer ownership of technology outcomes. Hiring priorities moved away from generic engineering roles towards profiles with deep platform and architectural expertise. The same approach supported early gen AI pilots by focusing human talent on high-value design and oversight rather than routine execution. The company later mapped critical skills across product management, full-stack engineering and design, supporting targeted development, career paths, reduced overlap and more skills-based hiring.
Vendor Partnerships Move Towards Outcomes
Companies also need to decide which capabilities they will deliberately build and sustain internally as agents perform more execution work. Effective CIOs take a skills-based view of the workforce, identify capabilities that differentiate each team, assess current proficiency and invest in development. Companies cannot delegate their understanding of AI’s effects on development road maps, platforms, data or architecture externally without losing control of value creation. They also should not build every capability internally, so the boundary needs clear definition and periodic reassessment as AI matures.
Vendor strategy is moving from managed services towards managed value. Outsourced development contracts increasingly link compensation to delivery speed, quality, modernisation progress or business impact. Older contracts rewarded activity, head count or volume. In an agentic environment, that model becomes costly because smaller groups of experienced engineers can guide faster and more flexible development. Companies increasingly expect vendors to use automation and AI tools rather than operate within rigid scope definitions.
Software-as-a-service models are also changing as vendors move from seat-based consumption towards API-first, outcome-linked software and agentic workflows. A bank that moved to a product platform operating model aligned vendor decisions with product road maps, introduced a central KPI for outcomes and negotiated AI-driven automation targets with every strategic vendor. Internal and external teams achieved a 10% productivity uplift.
Agentic AI links hiring, capability building and vendor strategy more tightly than before. Hiring determines where human judgement sits in the organisation. Capability development determines whether AI amplifies that judgement. Vendor relationships determine whether productivity gains stay with the enterprise or disappear into outdated delivery models. CIOs need to reassess hiring plans, invest in internal skills and reset vendor relationships around outcomes rather than effort. Close collaboration with business leaders remains essential. Organisations that delay may remain locked into structures that limit flexibility and future returns.
Source: McKinsey & Co
Image Credit: iStock
References:
McKinsey & Company (2026) Designing an end-to-end technology workforce for the AI-first era. S.l.: McKinsey & Company.
