Despite significant investment in digital laboratory technology, many scientists remain frustrated by fragmented workflows and disconnected data, according to a recent survey.
Here, Dara McCreary, Senior Scientific Business Analyst at Sapio Sciences, discusses the findings, and what the technology shortcomings mean for diagnostic laboratories seeking to improve efficiency while adopting AI responsibly.
What stands out most to you about current sentiment across laboratory professionals regarding electronic lab notebooks (ELNs) and AI tools?
In clinical laboratories, turnaround time can directly affect patient care. Our research found strong interest in AI-enabled tools, with 97 percent of scientists reporting that AI-powered ELNs could improve efficiency. However, many existing ELNs remain focused on documentation rather than helping scientists interpret results or determine next steps.
There is also understandable caution around AI in diagnostic settings. While AI can support tasks such as data interpretation, calculations, and workflow management, diagnostic decisions should remain under the oversight of qualified healthcare professionals.
Many labs still struggle with slow throughput and fragmented workflows. How do respondents describe the biggest day-to-day bottlenecks that ELNs and digital tools are meant to address?
Disconnected systems and fragmented data remain major barriers. In some laboratories, information is spread across numerous laboratory information management systems, making it difficult to bring together the data needed for interpretation and decision-making.
Our survey found that 51 percent of scientists spend too much time moving data between ELNs and other systems. Many still rely on spreadsheets to bridge gaps between platforms, creating additional opportunities for duplication and inconsistency.
The impact extends beyond operational efficiency. For example, a genetic variant identified in one laboratory may be relevant to findings generated elsewhere, but disconnected systems can make those links difficult to identify. In clinical diagnostics, fragmented data can delay interpretation and complicate patient care.
From a diagnostic perspective, where do scientists feel ELNs are genuinely improving efficiency, and where are they falling short of expectations?
ELNs have delivered clear benefits in documentation, traceability, and audit readiness. These capabilities are particularly valuable in regulated environments where laboratories must maintain comprehensive records for quality management and compliance purposes.
Where many ELNs fall short is in supporting interpretation and decision-making. They record what was done but often fail to capture why decisions were made. When key reasoning exists only in emails, spreadsheets, or conversations, reconstructing the path from sample to result becomes more difficult.
Laboratories increasingly need systems that capture both the underlying data and the rationale behind interpretations within the same auditable workflow.
What did the survey reveal about how comfortable scientists are with AI-assisted tools, particularly in data analysis, documentation, or decision support?
Scientists are already using AI in various aspects of laboratory work, but adoption within diagnostic workflows remains relatively limited.
Concerns around patient safety, accountability, and validation continue to influence adoption. At the same time, scientists are not rejecting AI outright. Instead, they want transparency. Our survey found that 81 percent of respondents would trust AI-generated recommendations only if they could review the underlying evidence and scientific rationale.
These findings suggest that explainability and human oversight will be essential as AI becomes more integrated into diagnostic workflows.
Were there differences in attitudes toward ELNs and AI between research-focused labs and clinically oriented environments?
The challenges are similar across research and clinical laboratories, but the consequences differ.
In research environments, inefficiencies can increase costs and slow discovery. In clinical settings, delays and uncertainty can affect diagnoses and treatment decisions. As a result, clinical laboratories tend to adopt new technologies more cautiously.
Both environments share a need for clear, defensible records that link data, analysis, and reported results. Maintaining that chain of evidence is critical for quality, reproducibility, and confidence in the final outcome.
Workflow integration is often cited as a barrier to adoption. How did respondents describe the impact of poorly integrated systems on productivity, data quality, and turnaround time?
Many diagnostic workflows depend on handoffs between teams and systems. Our research found that only five percent of scientists can analyze experimental data without support, highlighting the reliance on specialist expertise.
Next-generation sequencing is a common example. Laboratory teams may generate sequencing data, while bioinformatics teams perform analysis and validation. When multiple systems are involved, samples can spend valuable time waiting in queues before interpretation begins.
Each handoff introduces opportunities for delay, data fragmentation, and gaps in documentation. Linking analytical results directly to sample information, protocols, and supporting records can help improve efficiency and traceability throughout the workflow.
What concerns did scientists raise around data integrity, validation, or over-reliance on AI within ELNs?
Trust in diagnostic results extends beyond scientific accuracy. Laboratories must also consider regulatory requirements, quality standards, and accountability.
Data privacy remains a major concern, particularly when AI systems interact with sensitive patient information. Scientists want greater transparency around how data are handled, how models are developed, and how outputs are generated.
For AI to be adopted responsibly, laboratories need clear oversight and documentation of both human and AI contributions. The methods used, data sources consulted, and any uncertainties should remain visible and auditable.
Based on the survey results, what practical steps could labs take now to improve adoption and ensure ELNs and AI tools actually support diagnostic efficiency rather than add friction?
The first step is establishing a reliable data foundation. Many laboratories still work with data that are scattered across multiple systems, inconsistently formatted, or incompletely documented. Data quality and standardization must be addressed before AI can be implemented effectively.
AI should then be integrated into existing laboratory workflows rather than introduced as another standalone tool. The goal is to support data interpretation and decision-making while working within established quality and compliance frameworks.
Finally, laboratories should involve regulators early in the process. Regulatory engagement can help identify potential risks, clarify expectations, and ensure that AI-enabled workflows are developed in a way that is both effective and compliant.
Laboratories that focus on data quality, workflow integration, transparency, and governance will be best positioned to evaluate and implement AI in clinical practice.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
