How accurate do you think the “95% of AI projects never leave pilot” claim really is and are these failures mainly about the tech, or about deeper organisational issues?
It’s overwhelmingly an organisational issue. Most enterprises don’t struggle because the technology isn’t capable but because they approach AI with the wrong mindset. Too often, AI is treated as a fashionable investment or a way to speed up what the organisation already does. Leaders fixate on efficiency when the most important question is actually, “What can AI help us do that was never possible before?” That shift in mindset transforms AI from a tool into a genuine capability.
This becomes even more important as AI moves into business-critical functions. Over the past year, many organisations focused heavily on AI for IT operations, but the fastest growth over the next three years will be in areas like R&D, compliance and ESG reporting. These are domains where data volume, complexity and regulatory pressure are far higher. These functions already understand the value AI can unlock, yet they’re rarely given leadership of AI programmes. Instead, most organisations still assign the responsibility to technical teams alone, which is a missed opportunity. Business teams are the ones who live the workflows every day; they know where the friction is and what “good” looks like. Without them shaping strategy, it’s very difficult to scale beyond proof-of-concept.
So, if organisations hope to move beyond the 95% failure rate and close that execution gap, AI must become an executive priority.
How does DXC’s AdvisoryX Group help organisations turn isolated AI pilots into real, enterprise-wide transformation?
A lot of organisations have bold ambitions for AI, yet many find themselves stuck in a cycle of disconnected pilots that never quite translate into meaningful change. We recently launched AdvisoryX, a global advisory and consulting group designed to help enterprises address their most complex strategic, operational, and technology challenges.
What makes AdvisoryX Group different is the way it brings together the full breadth of DXC’s expertise. We combine corporate strategy, operational excellence, technology transformation, people and culture, finance and risk, and user experience into a single, integrated advisory model. That means clients aren’t just handed a roadmap; they work with teams who understand both the business ambition and the engineering realities required to make that ambition executable.
AdvisoryX blends consulting-led engagement with DXC’s deep engineering heritage, helping organisations diagnose the root causes of their challenges, design future operating models that reflect where the business is heading, and then execute those transformations at scale.
What does DXC’s Xponential framework look like in practice, and how much disruption should leaders expect when rolling out these AI capabilities?
The truth is, meaningful AI adoption does change the way an organisation operates. But with the right methodology, the change is structured and constructive.
Xponential covers the full AI lifecycle, starting with the foundation and ending with long-term production management. With five core pillars, this repeatable blueprint enables leaders to deploy AI with speed, quality and scale.
AI Core is where the groundwork happens. This is the data architecture, the modelling environment and the governance controls that make enterprise-grade AI possible. It’s the equivalent of laying the wiring before you switch on the lights.
AI Reinvent is where organisations start seeing real value. Here we apply AI to carefully chosen use cases, whether that’s human-assisted decision-making, semi-autonomous tasks or fully AI-native processes.
AI Interact which follows, is particularly transformative for employees. It reimagines human–AI collaboration by redesigning traditional processes into digital workflows that AI can understand. Instead of paper documents or manual steps, you create digital artefacts that users can interrogate or receive guidance on, which fundamentally changes how work gets done.
AI Validate and AI Manage keep everything on track. As AI becomes more agentic, continuous testing, quality control and human oversight become essential. And once systems are running in production, you need people who understand both the technology and the business context to keep performance, reliability and observability aligned. This is where many AI projects fail, so it’s essential and also where Xponential brings discipline.
Does Xponential’s fast pace raise risks around compliance, governance, or employee resistance and how do you mitigate them?
Speed only creates risk when governance is an afterthought. But with our approach to AI transformation, governance is built in from the start. Compliance, transparency and responsible AI principles are part of the foundation, not an add-on.
Of course, governance alone isn’t enough. AI transformation fails when leaders think of it as a technical rollout rather than an organisational redesign. If you try to introduce AI without rethinking the processes and roles around it, you’ll inevitably face resistance. People resist what they don’t understand or can’t see themselves in.
Our approach is deliberately human-centred. It promotes a Human + AI model where AI amplifies capability rather than displacing it. That demands new governance models, clearer decision rights and investment in skills. Ultimately, AI creates new workflows that require new skills. If leaders implement AI without redesigning the processes it will touch and without developing the workforce capable of operating in these new workflows, they’re unlikely to see the results they anticipate.
Can you share use-cases where this approach drove measurable results?
We always start with ourselves. DXC is our own customer zero, which means every AI solution we take to market has first been tested, refined and proven within our own environment. That internal early adoption accelerates learning and ensures what we deliver to clients is genuinely enterprise-ready. The pace is measured in days and weeks, not years.
For clients, we’re seeing strong outcomes across sectors. Ferrovial is a great example. They deployed more than 30 AI agents to streamline operational processes for around 24,000 employees. Those agents now support real-time decision-making and safer, more efficient operations at scale.
Another example is the European Space Agency. Through ASK ESA (an AI-powered platform that unifies and interrogates vast datasets) the agency has been able to accelerate research, enhance collaboration and vastly reduce the time required to generate insights. It’s a clear illustration of how governed, well-designed AI can unlock new capabilities in highly complex environments.
These are the kinds of transformations that happen when organisations move beyond isolated pilots and begin treating AI as a foundation for the way they operate, not an experiment happening in the margins.
Discussion about this post