A few years ago, I took on a digital transformation leadership role at a financial services company. The mandate sounded right. The title had weight. But within a few months it became clear that the mandate on paper and the mandate in practice were two very different things.
Every significant initiative needed someone else’s budget approval. Business unit heads were polite in meetings and quietly protective of their own agendas outside them. The decisions that actually shaped the customer experience, the product architecture, the data infrastructure, sat inside other functions. I had the responsibility to transform the business digitally but not the authority to move the parts that transformation required. I left within months, not out of frustration with the people, most of whom were genuinely capable, but because the structure of the role made meaningful change close to impossible to sustain.
The rise and fall of the Chief Digital Officer role over the past decade is a clear example of how weak role design can squander strong leadership without delivering meaningful impact. As we enter the AI era, we are on the brink of repeating the same mistake. Boards must define, with intent and precision, the mandate, decision rights, and accountability of the Chief AI Officer if they expect the role to create real value for the business.
The CDO angle
When many CDOs stepped into these roles, they brought heavyweight digital transformation credentials from technology and professional services firms, ambitious, commercially driven, and built to move fast. What they walked into were organisations whose structures had never been designed for fast at all.
The technology budget lived with the CIO, who had their own infrastructure priorities and annual planning cycles. The core business domains had technology, relationships and working practices built over decades that digital initiatives had to negotiate around rather than redesign. The business units, each with their own P&L pressures and leadership incentives, had little formal obligation to align with a transformation agenda that sometimes asked them to absorb short-term disruption for long-term gain.
IMD Business School found the average CDO tenure across industries sat at just 31 months, the shortest of any C-suite role. The World Economic Forum described a consistent arc: early enthusiasm followed by gradual disconnection, as the gap between responsibility and authority became impossible to bridge. Key executives would quietly disassociate. Parallel digital projects would spring up outside the CDO’s remit. The role would contract into an advisory function that nobody had formally decided it should become.
As one CDO put it in IMD’s research: “I’m not involved anymore to see the solution being implemented within the company. It will still be the digital baby that is growing up, but it’s not me or my team managing these implementations.”
We Saw Two Patterns, Two Outcomes
The CDO era produced enough evidence across industries to understand exactly what structural conditions determined success or failure. Reporting line mattered, but it was not the main variable. What mattered most was whether the CDO had genuine budget ownership, the authority to make decisions that affected other functions, and a mandate that had been accepted across the C-suite before they walked in the door. Where those three things were in place, transformation happened. Where they weren’t, the role became expensive theatre.
Here is a case where the mandate was BORROWED
A large North American financial institution hired a well-regarded CDO in the mid-2010s with a public commitment to digital transformation. The role reported to the CEO on paper. In practice, the technology budget sat entirely within the CIO’s function. Every digital initiative of any scale required a business case to be approved through IT planning cycles that ran on annual timelines. The CDO’s team could design the strategy, but had no authority to fund execution without negotiating with functions that had their own competing priorities and their own interpretations of what digital actually meant for their domain.
Within two years, the most commercially significant digital projects had either stalled in approval cycles or been quietly absorbed into the IT roadmap, where they were sequenced according to infrastructure priorities rather than customer impact. And the result – the CDO departed. The bank reabsorbed the function into its technology organisation and announced a refreshed digital strategy under the CIO’s leadership. The transformation it had promised its board and investors took several more years longer than originally committed, at significantly higher cost.
A European retail bank of comparable scale in a similar period tried to achieve the same outcome but structured the role differently. The CDO reported to the CEO but the budget came through an annual process controlled by finance and IT. Business units engaged selectively, on initiatives they had already decided to support, and deflected on anything that required them to change how they operated. The CDO spent the majority of their tenure making the case for decisions rather than implementing them. Three years in, the bank announced a restructure that eliminated the standalone role and distributed digital ownership back to the lines of business. The board described it as maturity. In practice it was a retreat to the structures that had made the CDO necessary in the first place.
The failure was not a strategy failure. The strategy was sound. It was a governance failure, and it was visible from the structural design of the role before the first hire was made.
Here is the other case, where the mandate was OWNED
A major North American media and retail group facing significant disruption from digital-native competitors took a fundamentally different approach. When they redesigned the CDO mandate in the late 2010s, the role came with its own ring-fenced transformation budget, sized to the ambition rather than constrained by IT’s existing allocation. The CDO had formal authority to override business unit decisions where they conflicted with the transformation programme, with escalation rights that went directly to the board, not through a technology or commercial intermediary. Business unit leaders understood, before the CDO arrived, that their cooperation was not optional.
Within three years the organisation had overhauled its customer-facing digital infrastructure, launched a data platform that enabled personalisation at scale, and integrated digital revenue streams that had previously been managed as separate experiments. The CDO did not leave after 31 months. The role evolved, as transformation roles should, into something more deeply embedded in how the business was run.
What made it work was not the reporting line alone. It was the combination of owned budget, clear decision rights, cross-functional authority that had been accepted at board level before the hire, and accountability for commercial outcomes rather than programme milestones.
Chief AI Officer – it risks the same issues as CDOs
IBM’s Institute for Business Value found that 26% of organisations had a Chief AI Officer in 2025. By 2026, that figure had reached 76%. That growth reflects genuine board-level intent. AI is being treated as a strategic priority that warrants executive ownership and not just an IT programme.
The structural risk is that the CAIO mandate is being built on the same framework that produced the CDO outcomes boards weren’t satisfied with, a broad remit, a visible title, and a budget and authority structure that quietly makes delivery impossible at pace.
Research in 2026 is already identifying CAIO appointments that lack budget authority and cross-functional decision-making power. The failure drivers are identical to those the CDO era surfaced: unclear mandate, insufficient budget, no formal standing across business units, success measures focused on activity rather than outcomes. Reading those descriptions, I recognise my own experience in financial services, almost word for word.
The Questions Boards Need to Answer Before the Hire
The CDO era’s central lesson is that the structural decisions made around the role matter more than the appointment itself. These are the questions that deserve board-level attention.
Does the CAIO own a budget, or will they be dependent? This is the single most important structural question and the one most often left unresolved. A CAIO who must build a business case and seek approval for every significant initiative through IT or finance planning cycles will move at those cycles’ pace, which is not the pace AI requires. If the board believes AI is a strategic priority, the transformation budget needs to reflect that structurally, ring-fenced, sized to the ambition, and owned by the CAIO with P&L accountability attached.
A CAIO can only make real decisions without constant escalation when three foundations are in place: a clear business case, proven CAIO capabilities, and a trusted ecosystem of partners. Start by defining where AI will move the needle on concrete business KPIs (for example, customer retention, unit cost, or revenue uplift), and build a quantified business case for each domain. Then match those domains to the CAIO’s track record and the specific capabilities of chosen technology and delivery partners, so the board and CXOs can see who is accountable for what, and why that accountability is credible. In that context, the CAIO should own end-to-end decision rights for AI within those domains including the impacted business KPIs as commits, while business unit leaders formally co-own the same KPIs and commit resources to execution.
For example, a bank might define a three-year AI programme to reduce retail credit losses by 15% and improve approval turnaround time by 30%. The CAIO is given direct ownership of these KPIs for all AI-driven credit decisions across products, backed by an in-house data science team and a specialist risk-analytics partner. The retail banking head and CRO co-own the same KPIs and sign off on an operating model where: (1) the CAIO approves teh program, model architectures, data pipelines, and deployment approaches without further escalation; (2) business leads provide frontline staff, process changes, and training to embed the AI into underwriting; and (3) performance against the shared KPIs is reviewed monthly. In this setup, the CAIO genuinely holds AI decision authority, and CXOs can support those decisions with confidence because they are grounded in a transparent business case, clear capabilities, and shared accountability for results.
The CDO failures that recur most often from various research is not failures of CEO intent. They are failures of C-suite acceptance. The CIO, CFO, and business unit heads who engaged selectively with the CDO based on their own priorities will do the same with the CAIO unless the board has made clear that selective engagement is not the option. Alignment at that level has to be a governance decision.
Historically, CDO roles became associated in many organisations with activity metrics, initiatives launched, frameworks delivered, assessments completed. Boards need to define, from the start, what commercial results the CAIO is accountable for; revenue enabled. cost reduced. risk identified and quantified. That accountability is more demanding to design upfront, but it is what gives the role genuine weight inside the organisation over time.
The Harder Point
There is a version of the CAIO appointment that is largely about signal, showing investors, employees and the market that AI is being taken seriously. That signal has value. But signal without structure produces CDO outcomes, and most boards have seen enough of those to know what they cost, in time, in money, in talent that moves on, and in competitive ground lost while structural questions remain unresolved.
My experience in financial services was not unusual. Variations of it played out across industries throughout the CDO era. Talented people, genuine board intent, and structures that made sustained delivery close to impossible. The structural gap between what the role was asked to do and what it was actually empowered to do was the consistent variable, not the quality of the people holding it.
The Chief AI Officer role has a genuine opportunity to be something the CDO role often couldn’t be. Whether it reaches that potential depends on the decisions boards make before the hire: about budget ownership, decision rights, cross-functional authority, and accountability for outcomes rather than activity.
AI is moving faster than digital did. The window for getting those decisions right is shorter than most boards appreciate. Think now, think again.
Leave a comment