As AI sweeps through the consulting industry and its clients, the traditional models used to deliver projects are no longer fit for purpose. This shift, especially in the Middle East where AI ambitions are accelerating rapidly, demands for a new consulting model, writes Endava executive Zane Ulhaq.
For the past two years, we’ve repeatedly heard that the overwhelming majority of generative AI initiatives never make it beyond the pilot stage. The industry has already spent enough time debating the causes, from poor data quality and immature governance, to unrealistic expectations and skills shortages.
There is truth in all of these explanations, but they share one underlying assumption. They treat AI project failure as a technology problem, or at best an organisational one. Far less attention has been given to the way AI itself is delivered, and whether the consulting and commercial models that have served enterprise technology for decades are still suited to a technology that evolves almost continuously.
That question is becoming increasingly relevant across the Middle East. Governments are accelerating AI adoption at unprecedented speed, with the UAE recently directing federal entities to deploy AI-powered services within 90 days, while Saudi Arabia continues to invest heavily in sovereign AI capabilities through initiatives such as HUMAIN.
Yet despite this momentum, many organisations are still procuring and delivering AI as though it were another large-scale transformation programme with clearly defined requirements, predictable milestones and a fixed destination.
The reality is very different. The AI era is not simply changing the technology organisations adopt. It is changing the amount of uncertainty they must be prepared to manage, and that has important implications for the relationship between clients and the partners helping them navigate that journey.
AI has changed the rules of delivery
Most enterprise transformation programmes follow a clear delivery path. Define the desired outcome, agree the scope, build the solution and measure success against the original plan. While no major programme is ever entirely predictable, the objective itself usually remains stable throughout delivery.
AI challenges that assumption. Models continue to improve soon after deployment begins. Regulation is evolving rapidly, particularly in sectors where governance and data residency are strategic priorities. At the same time, organisations are still learning how employees interact with AI in practice, often discovering that the greatest opportunities and challenges emerge only after solutions are placed into everyday use.
This is why AI increasingly behaves less like a traditional technology implementation and more like an ongoing capability that must be refined over time. The industry has been moving towards iterative delivery for years, and many organisations, including Endava, have long advocated for continuous innovation over large, multi-year transformation programmes. AI simply makes that approach significantly more important because the environment itself refuses to stand still.
The consequence is that certainty can no longer be treated as a prerequisite for delivery. Instead, organisations need delivery models that acknowledge uncertainty from the outset and provide the flexibility to adapt as technology, regulation and business priorities inevitably evolve.
The commercial model has become part of the delivery model
This shift raises an important question about how AI programmes are funded, governed and measured. Traditional consulting projects were built around a world where the majority of uncertainty could be removed before implementation began, and consequently, commercial success was linked to delivering against agreed milestones.
That approach becomes much harder to sustain when the destination itself continues to evolve. If organisations accept that AI programmes will require continuous refinement, then commercial models should evolve in the same direction. Rather than committing significant investment upfront, programmes can be structured around phased delivery, with each stage validating assumptions, demonstrating measurable business value and informing what comes next. Elements of commercial reward can increasingly be linked to agreed outcomes rather than simply the completion of predefined activities.
This is where shared accountability naturally develops into shared risk. Both client and partner remain invested in achieving meaningful business outcomes, creating stronger incentives to challenge assumptions early, adjust priorities when circumstances change and focus investment where value is being realised. In many respects, shared risk becomes another layer of governance, encouraging disciplined decision-making at precisely the time organisations are operating in their most uncertain environment.
Better incentives create better outcomes
Perhaps the greatest benefit of this approach is not commercial at all, but behavioural. Delivery models inevitably shape how both clients and consulting partners make decisions, often in ways that go unnoticed until projects begin to drift.
Traditional programmes can encourage organisations to define broad scopes early, attempting to anticipate every future requirement before meaningful learning has taken place. Consulting partners, in turn, are incentivised to deliver against those agreed commitments as efficiently as possible, even as priorities evolve around them. Neither side is necessarily acting incorrectly, but both are responding to the commercial framework that governs the engagement.
Shared accountability encourages a different dynamic. Organisations become more deliberate about prioritising use cases, defining measurable outcomes and maintaining executive ownership throughout delivery. Consulting partners remain focused not simply on implementation, but on ensuring each phase delivers tangible value before moving to the next. The result is a relationship built around continuous validation rather than periodic review, creating space for ambition, while introducing the discipline needed to ensure that ambition remains grounded in business outcomes.
The AI advantage
One of the defining characteristics of AI is the speed at which competitive advantage can disappear. Capabilities that appear differentiated today often become widely accessible within mere months. The rapid evolution of Abu Dhabi’s Falcon models illustrates this well, with successive generations arriving in quick succession as the technology continues to mature.
For organisations, this means AI should no longer be viewed as a sequence of standalone projects that begin and end with implementation. Competitive advantage will increasingly come from the ability to refine, improve and adapt continuously as both the technology and the business environment evolve.
That ultimately requires a different kind of partnership. One built less around delivering a fixed project and more around sustaining progress over time, with both client and consultancy aligned and invested in the same outcomes.
About the author: Zane Ulhaq is Head of MENA at Endava, a global digital transformation and IT consultancy.