When pharmaceutical facilities can cost hundreds of millions of euros, getting the numbers right from the outset is critical. Boehringer Ingelheim explains how it is combining AI, historical project data and robust governance to improve early-stage cost estimation and reduce investment risk.
Shutterstock - azrin_aziri
A new pharmaceutical plant can run to hundreds of millions of euros – and the decision to build it often has to be made before the design is anywhere near final. GMP requirements, highly specialised equipment, and relentless time-to-market pressure leave little room for guesswork. So how can a company estimate the cost and risk of a facility that, in effect, doesn't fully exist yet? That is the question that the pharmaceutical company Boehringer Ingelheim answered with its CapEx cost estimation framework. It combines standardised governance, data driven validation and AI supported data extraction to deliver reliable early cost estimates that enable faster, more confident investment decisions.
The pharmaceutical company Boehringer Ingelheim manages a growing portfolio of capital investment projects. As a global company it invests in a wide range of facilities that extend beyond pharmaceutical production. These facilities include, for example, plants to produce specialty chemicals for active pharmaceutical ingredients. They also include secondary production facilities, warehouses, and laboratories. For every investment the management has to assess the fundamental question of whether a facility is economically viable and what it means to invest in the U.S., China, or Europe.
Because project implementation models vary worldwide, differences in pricing structures and labour costs make cost estimation complex. Furthermore, once clinical trials have been successfully completed, investment has to be decided fast, before the final production process design is set. The objective was therefore to improve cost accuracy, provide reliable estimates earlier in the decision-making process, and leverage historical project data.
Implementation of a stage gate process
Building on more than a century of capital project experience, Boehringer Ingelheim established cost estimation as a dedicated governance discipline within its global CapEx organisation – alongside project controls, planning/scheduling, and contracts and claims management. This function is now fully integrated into the company’s Front-End Loading (FEL) system, with stage gates that govern project continuity and financing.
"Crucial decisions need to be made at a very early stage of projects. After all, in some cases, we're talking about investments in the two- to three-digit million range," explains Dr. Ralf Kehlenbeck, global head of project management office. "Early in an investment project changes can be implemented with relatively little effort and minimal cost impact. But as a rule, the later changes and decisions are made, the more dramatically costs rise. Therefore, we have now five gates in our process, of which one is related to scope freeze and one to design freeze."
Typically, a project receives full funding after the design freeze – relatively early in its lifecycle. From that point the plans are locked, and the work moves into detailed engineering, construction, commissioning, and qualification. "In the past, without a reliable scope and design freeze we had issues as especially big projects exceeded both schedule and budget targets," adds Kehlenbeck.
Today, Boehringer Ingelheim goes one step further: the final estimate validation is now based on quantitative metrics for each project estimate, benchmarked against a normalised basis – either comparable as-built projects internally or analogous as-built projects from industry peers.
"We have joined an industry benchmarking group for the life sciences sector where we exchange insights with our peers to ensure that construction costs and investment decisions align with what our colleagues in other parts of the world are implementing for similar projects. In other words, we maintain an open dialogue with industry peers," says Dr. Mirela Tsagkari, senior manager global CapEx Estimation Boehringer Ingelheim. "We are convinced that through collaboration, we can make even greater progress and manage appropriate capital expenditures and costs in the pharmaceutical sector more efficiently."
Boehringer Ingelheim is taking the lead to establish a standard – contributing to a broader industry conversation on cost engineering best practices.
Data-driven decision-making
A central goal was to build lasting transparency, consistency, and institutional knowledge across the organisation, for instance through cross-database benchmarks. Step by step, the company accumulates this distributed knowledge and consolidates its historical cost data – drawing on its ERP system and project document management system. A cloud-based data warehouse turns project, equipment, building, and services data into a lasting corporate asset, one that surfaces insights and informs future cost forecasts.
"We are moving away from intuition-based assessments toward data-driven decisions. I think the key here is to use historical data to determine the direction we want to take," explains Kehlenbeck.
A key advantage of the new framework is that Boehringer Ingelheim now runs cost estimation in-house – no longer relying on figures supplied by external contractors.
Role of AI in cost estimation
Capital projects generate vast amounts of documentation – engineering specs, procurement records, contracts. Reviewing all of it manually to extract the data cost estimation relies on is highly time-consuming, which is why Boehringer Ingelheim turned to AI-powered tools.
"Today AI is primarily used as a tool to save time – especially when it comes to finding the right information, extracting it, and converting it into a format that makes sense for cost estimation and benchmarking. Data extraction supported by LLMs has demonstrated time savings of about 20%, including verification activities, compared with manual extraction methods," says Tsagkari. "We have made significant progress in the development of AI agents to extract this information for the various departments."
AI is a companion in this process and will continue to evolve in the future. But AI is not a panacea. It can neither solve nor replace engineering. A great deal of engineering work is still required to link all these systems together. Human oversight remains essential: someone always needs to check whether the AI extracted the right information – and understand exactly what went wrong when it did not.
Results
With standardised governance, structured estimate validation, and an AI-supported data foundation in place, cost estimation has become an in-house core competency at Boehringer Ingelheim rather than a capability bought in from contractors. The payoff is greater transparency and closer collaboration between engineering, sourcing, finance, and the business functions. This enables faster, more confident decisions. With reliable cost estimates available early, Boehringer Ingelheim can move when it matters.
Because more reliable estimates are now available early in a project, the company can decide faster and compete more effectively. Beyond the individual project, the framework feeds Boehringer Ingelheim's wider digital transformation: clearer lifecycle cost visibility, stronger benchmarking, and a firmer basis for long-term asset planning.