Advances in artificial intelligence and advanced analytics are today enabling decision problems of increasing complexity to be tackled with greater precision and efficiency. Although virtually all large Spanish companies acknowledge room for improvement in costs and operational efficiency, 84.5% of complex decisions are still made primarily based on experience and the judgment of their teams, or with the support of basic tools such as spreadsheets, fixed rules, and traditional analytics solutions.
The First Barometer of Mathematical Optimization in Spain and Portugal, conducted by DECIDE | Linkroad, delves into the segment of companies that have not yet adopted prescriptive AI, the technology that not only predicts scenarios but also identifies the best possible decision among multiple alternatives, taking into account complex operational variables and constraints, in order to understand why they have not yet taken the plunge, and what they would need to do so.
The gap isn’t a lack of knowledge
One of the study’s most revealing findings is the contradiction between perception and action. Among the large Spanish companies that currently do not use prescriptive AI, 95% acknowledge that it would have some positive impact on their ROI. Of that percentage, 46% rate it as moderate or significant. Only 5% believe it would have no meaningful effect on their business.
This data dismisses the idea that the holdback is technological skepticism. The problem isn’t that organizations doubt the value of prescriptive AI, but that they have not managed, or have not prioritized, turning that conviction into action.
“Many companies already know that mathematical optimization works, they know it has impact, and in some cases they already know where to apply it. What’s lacking isn’t vision, it’s execution. And that can be solved with the right approach,” says Daniel Herrero, Global Capability Lead – Decision Intelligence at Linkroad.
Strategic priorities, the No. 1 brake
The study clearly identifies the barriers that explain this gap. The main one is strategic prioritization, cited by 19% of the companies that do not use it: there are other initiatives occupying the leadership’s focus, and prescriptive AI is pushed down on the agenda. They are followed, with 16% each, by the fact that this technology has not been contemplated within the organization to date, and by the lack of internally specialized talent to implement it.
Following these, the perceived solution complexity (14%), the difficulty in justifying ROI to management or investors (13%), and the lack of knowledge about what it is or how it actually works (11%) appear. Significantly, lack of budget is the last brake on the list, cited by only 10% of respondents, which debunks one of the most common arguments used to justify non-adoption.
“This ranking of barriers is very revealing. The investment isn’t the main problem. The problem is the agenda, the knowledge, and the internal capability. They’re obstacles that can be addressed with the right levers,” adds Daniel Herrero.
A low-risk pilot, the most effective lever to unlock adoption
The study also asked non-user companies what they would need to take the step. The most cited answer, at 26%, is: start with a low-risk pilot, which allows concrete results without compromising resources or critical processes.
They are followed, at 19% each, by the possibility that the solution won’t require a specialized internal technical team and that it integrates with current systems without major changes. Executive sponsorship and the clear quantification of economic impact sit at 18%, respectively.
These data point to a pattern. Companies are seeking concrete evidence, technical accessibility, and a gradual path toward adoption.
What would help better understand the real value of the technology?
In line with this, the study reveals that companies demand practical, tangible resources to help them understand the value of prescriptive AI before committing. Product demonstrations are the most valued resource, cited by 38% of respondents, followed by ROI calculators (37%) and success stories from the same sector (31%).
This demand profile reflects that the market doesn’t need more theoretical arguments. They need proofs, applied examples, tools that translate potential into decision-ready figures, and expert guidance during implementation.
“Adoption isn’t accelerated by more theory, but by evidence. When a company sees an applied case relevant to its reality, understands the economic impact, and has a team guiding its implementation, the jump from exploration to action becomes much more natural,” says Begoña López Piedra, CMO of DECIDE | Linkroad.
The future: mostly exploratory, but with growing openness
Despite the current barriers, the study detects a majority willingness to explore the technology in the coming years. 57% of companies not currently using prescriptive AI assign it an exploratory or experimental role within a two-to-three-year horizon. 26% see it as a potential point-in-time support for concrete decisions. Only 15% consider it not a fit for their operating model.
“Mathematical optimization already drives high-impact decisions across multiple sectors, with adoption continually accelerating. And the most natural path for organizations to adopt it begins with a well-designed pilot: a real problem, clear metrics, and executive backing,” concludes Duke Perrucci, CEO of Gurobi.
