Why AI Transformations Are Failing—And What CEOs Can Do
Summary
Enterprise AI initiatives are spending heavily but failing to deliver because leaders obsess over technology instead of treating change as a product that teams must choose and use. Phil Gilbert draws on IBM’s large-scale cultural transformation to outline a different approach: brand the change, build a dedicated cross-functional leadership team, treat intact teams as customers, and focus on adoption outcomes rather than technical enablement alone. Practical measures include team-specific onboarding, charging for participation to signal value, protecting and equipping middle managers, starting small, and scaling via proven advocates within the organisation.
Key Points
- Treat your AI programme as a high-value product: brand it, attach values and make it more than a technology push.
- Create a cross-functional, empowered leadership team accountable for adoption and measurable outcomes.
- Position intact teams as the customers — tailor entry points and let teams opt in or walk away.
- Measure outcomes (business impact) not inputs (training hours, generated text or code).
- Charge teams a fee or require budgeting for participation to create accountability and perceived value.
- Start small to expose cultural and system barriers, fix them, then scale deliberately.
- Identify “magic people” and equip middle managers as advocates to accelerate adoption.
- Cultural assimilation — not tech capability — is the primary barrier to lasting AI transformation.
Content summary
Gilbert argues that CEOs repeat past mistakes by privileging AI technology over the behavioural, organisational and operational changes required to use it. Using IBM’s “Hallmark” transformation as a template, he explains how branding, dedicated leadership, team-centred adoption, outcome metrics and financial skin-in-the-game turned a programme from an initiative into an organisation-wide capability.
The article lays out actionable steps: brand the change around shared values (avoid labels that narrow the remit), embed programme team members with early adopter teams to learn and iterate, require teams to budget for participation to surface commitment, focus manager training on how to manage teams using the new capabilities, and concentrate on replicating early wins via influential staffers.
Context and relevance
With billions being invested into enterprise AI and many projects underdelivering, this piece is highly relevant to CEOs, transformation leads and change teams. It reframes AI adoption from an IT or data science problem to a product and adoption challenge — aligning with broader trends emphasising people-centred digital transformation, behavioural design and outcome-driven metrics. If your organisation is paying for pilots that never scale, the tactics here help diagnose whether the barrier is technology, accountability, incentives or culture.
Why should I read this?
Short version — stop throwing money at models and start treating change like a product your teams actually want. If you’re leading an organisation and fed up with pilots that fizzle, this is the practical checklist you need: brand it, charge for it, embed help, measure outcomes and back the right managers and people. Quick, practical and directly usable.
Author’s style
Punchy and experience-led: Gilbert writes from large-scale, hands-on transformation work at IBM. If you care about making AI stick (not just launching it), his lessons amplify why culture and product thinking beat fresh tooling every time.
Source
Source: https://chiefexecutive.net/why-ai-transformations-are-failing-and-what-ceos-can-do/
