Electronic lab notebooks (ELNs) are widely used digital systems designed to document experiments, store data, and replace paper-based lab records. They are now standard infrastructure across biopharma R&D, clinical diagnostics, contract research organizations, and pharmaceutical manufacturing. However, new research from Sapio Sciences suggests ELNs often fall short of supporting how scientists actually work today.

Sapio Sciences conducted the study to assess scientists’ experiences with ELNs and AI tools in day-to-day laboratory workflows. The survey included 150 scientists working in US and European labs and examined ELN usability, data analysis capability, experiment reuse, and the growing use of shadow AI tools outside governed environments.

Only 62 percent of respondents said their ELNs allow them to work efficiently. Just five percent reported being able to analyze experimental results independently within their ELN, without relying on informatics specialists or data scientists.

Workflow rigidity and usability challenges persist

The study highlights several structural limitations that continue to affect ELNs. Workflow rigidity was a common concern. Only seven percent of scientists said their ELN could be adapted to new assays or experimental workflows without specialist support, limiting responsiveness as research programs evolve.

Usability issues were also prevalent. More than half of respondents said their ELNs are too complex and slow for day-to-day work. Configuration challenges compound the problem, with 71 percent saying ELNs are difficult to configure or adapt. Frustration was highest in pharmaceutical manufacturing environments, where standardized workflows often need frequent adjustment.

Manual data handling remains another burden. Fifty-one percent of scientists said they spend too much time importing and exporting data between ELNs and other systems, increasing the risk of errors and delays.

ELNs and experiment duplication

One of the most operationally significant findings relates to data reuse. Sixty-five percent of scientists said they have repeated experiments because previous results were difficult to find, interpret, or reuse within their ELNs.

For lab managers, experiment duplication has direct consequences. Repeating work consumes reagents, instrument time, and staff effort, while extending project timelines. When experimental data is technically stored but functionally inaccessible, the value of digital documentation is significantly reduced.

Lab Quality Management Certificate

The Lab Quality Management certificate is more than training—it’s a professional advantage.

Gain critical skills and IACET-approved CEUs that make a measurable difference.

The findings suggest that ELNs often succeed at capturing data but fail to return that data to scientists in usable, actionable ways.

ELN limitations are driving shadow AI use

As ELNs struggle to support analysis and interpretation, scientists are increasingly turning to external tools. The study found that 97 percent of respondents use some form of AI to support lab work, with 77 percent using public generative AI tools alongside their ELN.

Nearly half of respondents said they access public AI tools through personal accounts rather than company-managed systems. This shadow AI use reflects unmet analytical needs rather than a lack of awareness around data governance or security risks.

When ELNs cannot support exploratory analysis or help guide next steps, scientists look elsewhere for assistance, even when those tools sit outside validated and governed lab environments.

What scientists want from an AI lab notebook

When asked about future expectations, scientists emphasized that the next generation of ELNs should move beyond documentation. Ninety-six percent said future systems must help interpret data, not just capture it.

Demand for an AI lab notebook centers on interaction and guidance. Ninety-five percent want conversational, text-based interfaces, while 78 percent expressed interest in voice interaction. Scientists also want built-in, field-specific AI capabilities, including support for retrosynthesis, toxicity prediction, molecular binding simulations, and genetic sequence optimization.

Importantly, respondents emphasized that AI should augment scientific reasoning rather than replace it. Most said they would only trust AI-generated insights if they could review the underlying evidence and scientific logic.

Implications for lab managers and laboratory operations

For laboratory leaders, the findings highlight a growing gap between ELNs as infrastructure and ELNs as functional tools. Systems that cannot support analysis, reuse, and adaptive workflows create inefficiencies and encourage ungoverned workarounds.

As AI becomes more embedded in scientific work, the study suggests lab managers will need to reassess whether their ELNs can support governed AI capabilities or whether existing limitations are unintentionally driving shadow AI use. The shift toward an AI lab notebook reflects not a desire for automation, but a demand for tools that actively support modern scientific practice.

This article was created with the assistance of Generative AI and has undergone editorial review before publishing.