Elsevier survey: 63% of corporate R&D staff say a key decision rested on flawed data

More than three quarters (76%) of the R&D leaders, managers and researchers Elsevier surveyed say they are confident in the quality of their data. Yet 63% of those who answered say their organization has at some point made an important R&D decision using data or evidence later found to be inaccurate, incomplete or outdated. About half (51%) say their organization has had to make an important R&D decision before all the evidence was in.

Elsevier published the results Nov. 12 in its Corporate R&D of the Future report. The online survey ran by email invitation from August to October 2026 and drew 1,038 respondents at R&D-led corporations in 55 countries. The figures are weighted, and the flawed-decision figures exclude respondents who answered “don’t know” or preferred not to say. Elsevier also sells evidence and AI tools to corporate R&D teams. Its LeapSpace research workspace went live in January and added claim-checking tools in June.

While 78% of respondents say their teams can access the third-party evidence they need, 56% of that group still report at least one barrier. Across respondents, the leading obstacles were scattered sources (42%), slow finding and synthesis (32%) and uncertainty about what evidence exists (30%). “Fragmented data sources and disconnected systems create silos that slow down analysis and reduce transparency and weaken confidence in outcomes,” an R&D leader in Europe told the survey.

AI adds another layer to those evidence-quality concerns as R&D teams adopt tools that can generate inaccurate answers or unsupported claims. Nearly half of respondents (47%) say their organization’s R&D function uses AI regularly, and another 40% say it is piloting the technology. For literature searches, 48% of those answering use general-purpose tools such as ChatGPT or Copilot, compared with 23% using specialist scientific AI tools. Elsevier has a commercial stake in how companies address those risks: its LeapSpace research workspace includes claim-checking tools added in June.

Respondents’ priorities include improving how evidence feeds into decisions. Asked what their organizations should focus on over the next five years, 62% selected turning data and evidence into actionable insights, and 59% chose tailoring AI to organizational knowledge and workflows. Building AI skills across the organization was the most-selected priority, at 65%.

A Pistoia Alliance project offers a concrete example of how life-sciences companies are approaching reliability. Christian Baber, chief portfolio officer at the Pistoia Alliance, a nonprofit that promotes collaboration in life-sciences R&D, told R&D World in an Oct. 29 interview that the project covers large language models that build instrument data converters, which translate instrument output between formats. The models are “not consistent enough” to do that conversion directly, he said, so teams have them write a Python script instead. “Once that Python script is written, it’s testable and validatable,” Baber said. “You use something unpredictable to generate something predictable.” He added that the approach “is much more supported by regulators as well.”