The buck stops with you: scaling AI is a leadership choiceOwnership is often the missing piece. The proliferation of failed, small-scale AI pilots is a symptom of diffused responsibility. When AI is everyone’s job, it becomes nobody’s mandate.STU SCOTIS AND DAVID ALONSOAI pilots are useful for building awareness, but they are no substitute for strategyGift this article4 min read17 hours agoPilots are easy. Progress is not. Artificial intelligence doesn’t fail because the technology isn’t ready, it fails because leaders don’t commit. Most organisations spin up use cases that never scale. The few pulling ahead aren’t more technical, they’re more decisive.Leaders are noticing a pattern. According toa Deloitte survey of Australian CFOsconducted earlier this year, 80 per cent said AI was used in “pockets” across their organisation.Just 3 per cent said AI was used extensively.With only a small group of organisations implementing AI at scale, many still blame the technology for not being advanced enough.The problem isn’t the technology, it’s the leadership.When we published our firstPath to Scaleguide last year, we warned that simply launching and relaunching AI pilots and demos would do nothing to move the needle on productivity or boost the bottom line.That was when generative AI was the main focus. Now, with the arrival of agentic AI, the risk has doubled.Orphaned use cases will limp along, wasting time and money, until they are quietly retired. But the opportunity has also doubled, if leaders consider the points outlined in our updated Path to Scale 2.0.From use case hunting to ownershipOwnership is often the missing piece. The proliferation of failed, small-scale AI pilots is usually a symptom of diffused responsibility. When AI is everyone’s job, it becomes nobody’s mandate.Stu Scotis is Global Agentic AI Leader at DeloitteWithout direct accountability for AI transformation, difficult but necessary decisions, like pulling the plug on a pilot or committing the spend for a truly transformative project, won’t happen when they should.Organisations should appoint a single accountable executive and empower them to set priorities, resolve trade-offs and take on risk. AI should then sit permanently on the board agenda alongside cyber and safety, with directors and the accountable executive working together to build a strong governance framework.Many leaders see governance as a brake. In reality, strong guardrails create the operational trust that enables scale, safely and at speed. The arrival of agentic AI only makes this even more urgent. As autonomous AI agents begin to take on tasks with real responsibility, the consequences of poor governance could be catastrophic.Strategy with disciplineAI pilots are useful for building awareness, but they are no substitute for strategy. Strategy is the non-negotiables: what you fund, why, and for how long. To scale AI and deliver value, investments must target end-to-end processes that matter.Organisational politics will inevitably create funding demands from every business function. If an AI pilot is not geared towards reshaping a core value stream or end-to-end process, do not fund it.A “less is more” mindset should guide AI investment decisions. One pilot with proven results is worth more than 30 with no clear benefit. That said, a degree of well-placed ambition is healthy, especially given the transformative potential of agentic AI.We believe leaders should start by identifying three to five agent-driven initiatives closely tied to bottom-line outcomes such as increasing revenue, reducing costs, shortening cycle times or improving quality.David Alonso is National AI Market Lead at Deloitte AustraliaEach should be tested in short windows with clear targets. Scale what works. Stop what does not. Every initiative must have defined ROI metrics and undergo pilot validation before wider rollout, or your ambitions will never get off the ground.Lead with the change you expectCreating executive ownership and disciplined transformation strategies requires an organisation-wide cultural shift. This won’t happen through AI fluency training or internal communications alone. The tone must be set from the top.Leaders must role-model fluency by emphasising the outcomes of AI transformation and by aligning incentives to end-to-end results. They must also be prepared to own the consequences of success by bolstering change readiness.Agentic AI’s promise goes far beyond streamlining activities or cutting costs. At scale it will fundamentally change the composition of the workforce, with humans and agents working side-by-side. Clear autonomy thresholds must be defined so everyone understands when agents act alone, when validation is required and when collaboration is expected.AI is now table stakes. Waiting to see where it lands is not a strategy. If you’re not already on the path to scale, you must start now. Keep tinkering and you will stay in the trough. Own the transformation and you set the pace. Remember, AI will scale only as fast as leaders commit to move.Stu Scotis is Global Agentic AI Leader at Deloitte. David Alonso is National AI Market Lead at Deloitte Australia.Editor’s note: This is part one of a three-part article series.-DisclaimerThis publication contains general information only and Deloitte is not, by means of this publication, rendering accounting, business, financial, investment, legal, tax, or other professional advice or services. This publication is not a substitute for such professional advice or services, nor should it be used as a basis for any decision or action that may affect your business. Before making any decision or taking any action that may affect your business, you should consult a qualified professional adviser.Deloitte shall not be responsible for any loss sustained by any person who relies on this publication.About DeloitteDeloitte refers to one or more of Deloitte Touche Tohmatsu Limited, a UK private company limited by guarantee (“DTTL”), its network of member firms, and their related entities. DTTL and each of its member firms are legally separate and independent entities. Please seewww.deloitte.com/auto learn more.Copyright © 2025 Deloitte Development LLC. All rights reserved.-More CoverageHumans aren’t great at human skills, and that is a problemKatherine WannanBuilding future-ready AI infrastructure: how organisations can thrive in an AI-driven futureJason Hutchinson and James AllanThree ways AI can advance legacy tech modernisationTim Smith, Faruk Muratovic, Bill Briggs, and Diana Kearns-Manolatos