Lab Design: Connected Labs

A connected laboratory, where systems and devices communicate seamlessly, is the foundation to unlock AI’s full capabilities

Jeroen de Haas at Labforward

Artificial intelligence (AI) is revolutionising various industries, and scientific research is no exception. The application of AI in scientific research holds immense potential, offering new avenues for discovery, automation and efficiency. Yet, many laboratories struggle to leverage AI effectively due to fragmented systems and data silos. By integrating digital tools regardless of their vendor origin, laboratories can ensure that data flows freely across platforms, enabling AI to deliver accurate, actionable insights.

AI thrives on data. Specifically, large quantities of high-quality, well-structured data. In a typical research environment, data is generated from various sources, including instruments, sensors, information systems, workflows and human input. Without proper integration and connectivity, these data sources remain isolated, limiting the ability of applied AI technologies to be effective as they are missing context and metadata.

A connected laboratory eliminates these barriers by enabling interoperability between systems. Vendor-agnostic solutions create a foundation where AI can be deployed efficiently across diverse equipment and platforms. This ecosystem approach facilitates seamless data collection, analysis and management, which are crucial for any AI-driven strategy.

In a connected laboratory, AI has the potential to connect dots between equipment performance and experimental outcomes, analyse patterns across multiple experiments over time, predict potential issues before they cause major problems and provide actionable recommendations.

Imagine a system that monitors your thermal cycler’s performance and correlates it with experimental outcomes. When the system detects a pattern of subtle inconsistencies across 15 PCR runs in the past month, it automatically alerts the team, noting a 78% probability that Column 3 needs recalibration. By identifying this trend before it leads to experiment failure, the AI helps prevent wasted resources and timeline delays. Such predictive insights, only possible through the interconnection of equipment data, experimental results and maintenance logs, exemplify how a connected lab transforms scattered data points into actionable intelligence.

The ability to easily capture, organise and retrieve experimental data is critical for AI, which relies on well-structured data sets to generate meaningful insights. An electronic lab notebook (ELN) provides the core platform for collecting and managing experimental data. As a vendor-agnostic tool, it can integrate with various devices and software solutions, enabling researchers to document experiments digitally, in real time and in one place, providing the perfect foundation for AI-generated insights. An ELN not only ensures the traceability and integrity of experimental data collected across many platforms, but also allows for collaboration and sharing across teams, paving the way for AI to work on a larger scale. With all data centralised in one system, researchers can more readily apply AI algorithms to detect patterns, optimise experiments and even predict outcomes based on historical data.

While an ELN manages data at the documentation level, a laboratory execution system (LES) is the key to connecting devices and automating workflows across the lab, and therefore an essential tool of the connected lab (Figure 1). An LES that has a vendor-agnostic nature can interface with a wide range of laboratory instruments, enabling them to communicate with each other and with the central data system. By integrating an LES across vendors’ boundaries, labs can create automated workflows that generate consistent, high-quality data – exactly the type of data AI needs in order to learn and make predictions from. Automation not only reduces the chances of human error but also increases the volume of data generated, including more meta and process data, allowing AI models to be trained more efficiently. The LES thus acts as the bridge between physical lab activities and digital AI applications, ensuring that no data is lost in translation.

Figure 1: LES: Connecting devices for data-driven AI workflows

In addition, the modelling of digital processes in a LES enables AI-based optimisation of these processes as they are stored and developed in machine-readable and well-defined formats.

For example, when a high-performance liquid chromatography (HPLC) protocol is digitally modelled in the LES, AI can analyse vast amounts of chromatographic data to automatically optimise separation conditions. By incorporating real-time data from temperature and humidity sensors, the AI can suggest precise adjustments to maintain ideal environmental conditions, improving retention times and peak shapes. Additionally, the LES can send alarms and notifications if conditions deviate from optimal ranges, ensuring immediate corrective actions. This not only enhances the accuracy and reliability of the chromatography results, but also reduces the need for manual monitoring and adjustments, streamlining the overall analytical process.

“[The LES] not only enhances the accuracy and reliability of the chromatography results, but also reduces the need for manual monitoring and adjustments, streamlining the overall analytical process”

AI’s effectiveness in the lab also depends on real-time data availability. An AI-powered digital lab assistant addresses this challenge by allowing researchers to capture data hands-free without interrupting their experiments and automatically collate relevant metadata. This hands-free interaction ensures that all relevant information is recorded on the spot, providing AI with up-to-date data to analyse. An AI-powered digital lab assistant can integrate with other lab systems, such as an ELN, laboratory information management system (LIMS) and/or LES, as well as inventory management systems, making it a valuable component of a connected laboratory. Researchers can use voice commands to input data directly into the ELN without using their hands, keep track and manage their inventory, or even control devices – enhancing productivity while ensuring that AI systems have access to the most current data. This real-time hands-free capture is critical for AI applications that rely on time-sensitive data, such as process optimisation and predictive analytics.

For example, during a cell culture experiment, a scientist can use voice commands to log real-time observations about cell morphology, record pH measurements and adjust incubator settings, all while maintaining sterile technique at the biosafety cabinet. The digital lab assistant can not only capture this data but also provide immediate AI-driven feedback – such as alerting the researcher that the current growth rate deviates from historical patterns – and suggesting adjustments to media composition based on successful past experiments. Through natural conversation, the assistant can deliver these insights precisely when needed, allowing researchers to make data-driven decisions without breaking their workflow or removing their gloves.

Vendor lock-in remains one of the biggest obstacles for labs aiming to implement AI. Proprietary systems often fail to integrate smoothly with equipment and software from other vendors, creating data silos and limiting the scalability of AI applications. Creating a connected laboratory with vendor-agnostic tools offers several advantages when integrating AI:

• Data consistency and accessibility:

AI models require high-quality, well-organised data. A connected lab ensures that data from various instruments and systems is consistently captured, stored and accessible in a central location, ready for AI processing

• Scalability:

As AI applications evolve, they require increasing amounts of data and more complex workflows. A connected, vendor-agnostic lab can scale easily, accommodating additional instruments and data sources without the limitations imposed by proprietary systems

• Automation:

By linking devices and systems, a connected laboratory allows for the automation of data collection and workflow processes, leading to faster and more reliable AI-driven outcomes

• Enhanced collaboration:

Connectivity not only enables data to flow between systems, but also facilitates collaboration between researchers. Teams can share data across platforms, providing AI with a richer data set from which to generate insights

• Real-time insights:

AI systems thrive on real-time data, especially in areas like process optimisation and predictive modelling. The integration of tools like an AI-powered digital lab assistant ensures that data is continuously captured, giving AI the most current information to analyse.

Vendor-agnostic tools overcome these challenges by offering seamless integration with a wide range of instruments and software regardless of their vendor origin. This flexibility ensures that labs are not tied to a single provider, allowing them to choose the best AI and automation tools for their specific needs. In turn, this makes it easier to scale AI initiatives, expand capabilities and adopt new technologies as they emerge.

AI has the potential to transform laboratory operations, offering unprecedented insights and efficiencies. However, these benefits can only be fully realised in a connected laboratory where systems, data and devices are integrated and interoperable. API-friendly ELN, LIMS or LES, as well as AI-powered digital lab assistants, are key allies in running a connected and automated lab, and enabling labs to harness the full potential of AI. As the scientific community continues to embrace digital transformation, the importance of connectivity will only grow, making it a foundational element for AI-driven research and innovation.