For five years, the Niji Barometer has surveyed decision-makers at large French organizations: CIOs, CEOs, business leaders and heads of transformation. For this 5th edition, more than 100 decision-makers from mid-sized companies, SBF 120 and CAC 40 companies, and public administrations shared their challenges and trade-offs. Each year, the Barometer assesses organizational maturity across five pillars tracked since the first edition: strategic vision, governance, technology enablers, culture and skills, and performance measurement. This 2026 edition adds three new thematic focus areas: IT simplification, the impact of AI on Workforce Planning, and agentic AI platforms.
A double scissor effect
The first scissor effect shows up clearly in the numbers. On one side, 98% of organizations are maintaining or raising their expectations of technology's contribution to performance, a pressure that is not easing and has even grown from the 90% measured last year. On the other, 53% report flat or declining budgets, at a time when run, SaaS and cloud costs are, with few exceptions, on the rise. AI makes the equation worse: it opens the floodgates to new needs (use cases, data, platforms, APIs) and generates a wealth of projects that budgets cannot absorb. Doing much more with even less has become the rule.
But there is a second scissor effect, less visible and more dangerous. AI carries a real promise: producing faster, simplifying systems, freeing up considerable productivity margins. That promise is documented: organizations now project 40 to 60% gains over three years in their technology factories, compared with 15 to 20% a year ago. The problem is that the same AI, poorly orchestrated within an incomplete transformation, can produce the exact opposite: an extra layer of complexity, new projects with no governance, a bubble of technical debt on top of already fragile IT systems. The two blades combine: pressure is at its peak, room for maneuver is shrinking, and AI is both the main source of tension and the only way out.
Six figures that shed light on the picture
Beyond the double scissor effect, six Barometer indicators deserve particular attention, because they reveal the gap between stated ambition and actual transformation:
- 7 organizations out of 10 are already using AI to optimize their processes and IT. Adoption is broad.
- 92% vs. 19%: 92% use AI in their technology factories, but only 19% have deployed it systemically across their entire value chain. AI is present everywhere, but not yet transformative anywhere.
- 63% of organizations have a direct executive committee lead on AI: unprecedented sponsorship that shows how the topic has risen to the strategic level.
- 55% now require an ROI forecast before launching an AI project. The era of experimentation without accountability is over.
- Only 23% anticipate the impact of AI on Workforce Planning, even though that is precisely where the transformation will succeed or fail.
- 77% cite operational efficiency as their top technology priority, ahead of customer satisfaction and growth. AI must deliver measurable effects, not transformation for its own sake.
The 92%/19% contrast is the key figure of this edition: AI has been adopted on the surface by almost every organization, but mastered in depth by fewer than one in five. The competitiveness of the coming years will be decided in that gap.
IT modernization is not moving fast enough
Technology maturity is progressing. IT debt management has even made a remarkable leap this year. But this progress is still not enough to transform the technology ecosystem in depth, simplify architectures and free up the necessary budget headroom. Persistent blind spots are holding transformation back: redesigning cross-functional processes, change management, measuring the value generated, and preparing data to feed AI models. With no room to maneuver, responses remain largely short-term: demand gets triaged rather than operating models transformed.
The figure that illustrates this gap is stark: 92% of organizations use AI in their technology factories, but only 19% have deployed it systemically across their entire digital value chain. AI has been adopted on the surface, not yet anchored in depth. And human resources are not keeping up: only 23% of organizations anticipate the impact of AI on their jobs and skills, even though that is precisely where the transformation will succeed or fail.
The reinvestment paradox
This is the central knot of this edition. Organizations know that AI generates productivity gains: they project them at 40-60% over three years within their tech factories, an estimate that has more than doubled in a year. The executive committee is involved: 63% have a direct lead on AI. Budgets have been ring-fenced despite general restrictions.
And yet: only 33% reallocate part of these gains to transformation and to paying down technology debt. The remaining two thirds absorb the gains into new projects or hand them back to other budget lines. The result: AI creates local value, but does not fund systemic transformation. It may even make things worse, by generating ever more projects and run costs in IT systems that have not yet been simplified.
This is precisely the cycle that has to be broken for AI to become a lever of lasting transformation rather than a source of additional complexity.
Seven levers to transform from 2027
In this context, making information systems sustainable with AI is a safe strategy for reinvesting the gains: reallocating AI gains to the technology factory improves productivity, which in turn frees up new margins. The 2027 budget is the opportunity to set this dynamic in motion. Our recommendations fall into seven points:
- A gains redistribution policy - define, with top management, the share of AI productivity gains retained to fund the transformation of IT, rather than handing it all back or consuming it in new projects.
- A refactoring strategy - use AI to rewrite and simplify legacy systems and the most expensive SaaS products, replacing them with in-house solutions, to reduce run costs durably.
- A stronger regulatory framework for AI - a data usage policy, a cyber reference framework, an AI architecture framework, supplier qualification, and a structured run to oversee AI in production.
- An industrialized agentic platform - a shared foundation to accelerate use cases, a safeguard against the debt bubble, and a gradual alternative to some intermediate SaaS products.
- A transformation of the Target Operating Model - a new way of working, organizing and collaborating between humans and agents, end to end across the value chain.
- Structured Workforce Planning adapted to AI - well beyond one-off workshops: building new skills and new profiles, with a sourcing strategy as a lever to cushion the impact on headcount.
- An ambitious plan to transform tech factories - so that technology factories benefit fully from AI and generate gains that serve the business, IT and the refactoring of the IT landscape.
The full edition of the Niji 2026 Barometer, with detailed data, industry benchmark and action plan, is available on request. It lets you see exactly where you stand compared with the 100+ organizations in the panel, and identify the priority levers for your context. Get in touch below.
The question is no longer whether AI will transform your organization. It is already happening: at 92% in tech factories, at 63% at executive committee level. The real question is whether this transformation will be endured or steered. Whether the gains will be reinvested or absorbed. Whether the double scissor effect will cut you, or be cut through.
