From potential to results: how pharmaceutical companies measure the impact of AI on their forecasts
elglobalfarma.com
The global pharmaceutical industry is undergoing a quiet but profound transformation in the way it anticipates the future. A new report on forecasting and artificial intelligence (AI), produced by Evaluate Pharma, shows that compared to just a year ago, forecasting teams are operating in a much more demanding environment, where innovation is no longer enough as a promise. According to the document, the question pharmaceutical companies are asking themselves is not whether AI can help, but how it is being used to improve concrete results, increase efficiency, and support strategic decisions under pressure.
This shift in focus marks the transition from exploration to execution. Therefore, the consulting firm’s analysis emphasizes that the trends that will define pharmaceutical forecasting in 2026 do not revolve around grand futuristic visions, but rather practical implementation, sound governance, and the ability to “do more with less” without sacrificing quality.
Artificial intelligence continues to be a central topic of conversation, but its role has matured. After years of pilot projects, proof of concept tests, and debates about data preparation, he points out that organizations are beginning to integrate AI directly into their daily forecasting workflows. He also argues that today, the value of AI is no longer measured by its abstract potential, but by specific and repeatable use cases: automatic extrapolation of trends for brands in the market, rapid sensitivity tests on key assumptions, or machine learning models that help build baseline forecasts when there is sufficient historical data.
Far from replacing forecasters, AI is establishing itself as a mechanism for efficiency. It automates routine tasks, accelerates iteration cycles, and frees up expert time for critical analysis, interpretation, and strategic questioning. Human oversight remains essential, but now focuses on providing judgment, context, and experience, rather than manually feeding complex models. This shift also reflects a growing demand for transparency among peers. “Teams want to know what is working in practice, which solutions have been scaled, and where clear limits to automation still exist,” he says.
Along with the more pragmatic adoption of AI, the document identifies an emerging structural change in the organization of forecasting, especially for brands in the market. More and more companies are experimenting with centralized models, in which base forecasts are generated automatically from a global or regional center and then distributed to countries for review. In its most advanced form, this approach significantly reduces local ownership of the forecast. Teams in countries only intervene when there is a clear, evidence-based reason to deviate from the central scenario. The argument is compelling: greater efficiency, methodological consistency, and less dependence on scarce local resources.
This model represents a break with tradition. Historically, the report recalls that the early years after a product launch were characterized by highly granular, patient-based, locally managed forecasts. Therefore, it considers that moving toward automated approaches based on volume or sales in earlier stages of the life cycle represents a profound cultural change.
Another aspect analyzed in the report is adoption, which it admits is uneven. “Some organizations are moving forward gradually; others are firmly committed to centralization. The long-term impact on forecast quality, insight generation, and organizational agility remains to be seen, but the direction is clear: more automation and stronger central governance,” it notes.
Efficiency pressures
The document also mentions that the operational tensions identified in previous years not only persist but are intensifying. Forecasting teams face staff reductions, increasingly complex product portfolios, and growing scrutiny from senior management. “This context reinforces the drive toward simplification, automation, and selective outsourcing, and many companies are rethinking which activities really require in-depth internal knowledge and which can be standardized or supported by external expertise,” it states.
This is creating greater demand for flexible, on-demand support. Instead of expanding internal structures, companies are looking for specialists to cover work peaks, accelerate transformations, or fill specific capacity gaps. According to the report, the big question is whether this model will be sustainable in the long term or whether, with increasing complexity, teams will need to grow again. “The answer will depend largely on the ability of AI to absorb operational load without eroding forecast quality,” he adds.
On the other hand, he points out that there is a growing consensus that forecasting models have become very complex. In this context, he emphasizes that simplification promises greater transparency, better understanding by stakeholders, and more confidence in the results. However, he argues that putting this intention into practice remains one of the great challenges. In reality, models continue to accumulate assumptions, segments, and temporary “patches,” driven by inherited expectations, internal demands, or the fear of losing analytical rigor.
“Rather than widespread simplification, what we see is a gap between recognizing the problem and the ability to solve it. This opens the door to clearer modeling frameworks, better-defined principles, and governance that distinguishes when complexity adds value and when it only adds noise,” he says.
Likewise, cultural factors play a decisive role. In regions such as Asia-Pacific, differences in how uncertainty, detail, and validation are approached profoundly influence how forecasts are constructed and used. Recognizing these differences is no longer optional: it is a requirement for effective consolidation.
The pressure on forecasting teams comes not only from efficiency, but also from trust. “Senior managers want to understand not only the figures, but also the path that led to them,” he says. This has prompted renewed interest in forecasting audits, whether formal or informal, as a tool for identifying biases, structural weaknesses, and opportunities for improvement. Far from being a retrospective exercise, auditing is positioned as a forward-looking capability designed to raise standards, defend decisions, and strengthen credibility.
Looking ahead to this year, the document notes that pharmaceutical forecasting will become “more pragmatic, more centralized, and more accountable.” Although the transition from possibility to practice is underway, the real challenge is execution. “The teams that succeed will be those capable of balancing automation and human expertise, efficiency and insight, global consistency and local relevance. Because, in the end, forecasting is not just about producing a number, but about building a robust, adaptable system capable of withstanding scrutiny,” it concludes.
