Customize SciSure to work your way

AI has become one of the most talked-about technologies in science. The models are already demonstrating their promise to accelerate discovery, generate hypotheses, and uncover insights hidden in complex datasets​^1,2​. But translating that potential into day-to-day scientific workflows is still an evolving process.

One of the key challenges is that many AI tools do not yet have direct access to the data scientists work with every day. Experimental records sit inside ELNs, instruments generate data in separate systems, and operational information lives in yet another layer of software. To use AI, researchers frequently have to extract data manually and feed it into external tools. This approach quickly breaks down at scale.

For AI in labs to deliver real scientific value, it must move beyond standalone point solutions or chat interfaces, and become directly connected to the digital research environment. Recognizing this challenge, SciSure is introducing exiting new AI capabilities into the Scientific Management Platform (SMP), designed to allow intelligent agents to interact directly with scientific data and research workflows.

In this article, we explore why connecting AI to the digital lab is the critical next step for AI in labs, and how new integration approaches are beginning to unlock more powerful ways for scientists to search, interpret, and learn from their experimental data.

Why AI in labs has been slow to deliver practical value

Despite rapid advances in AI models, applying AI effectively within scientific environments is still developing, particularly when it comes to integrating AI with complex research data and workflows. The issue is not a lack of interest or imagination among researchers. Instead, progress is shaped by the realities of how scientific data is generated, stored, and accessed.

Historically, labs have been slower than many other industries to adopt digital infrastructure. Even today, some research environments still rely on paper notebooks or fragmented digital tools to document experimental work. While the transition to ELNs and integrated data platforms is accelerating, many organizations are still navigating the shift from analog to fully digital research workflows.  

At the same time, the scale of scientific data has grown dramatically. Modern research generates enormous volumes of experimental results, instrument outputs, and analytical datasets; far more than could ever be captured in traditional lab documentation. This explosion of data has made digital platforms essential for managing research activities, but it has also created new challenges for applying AI.

Most AI tools available today operate as standalone interfaces. Scientists interact with them through chat windows or external applications, manually uploading documents or datasets when they want to analyze specific information. While this approach can be useful for small tasks, it quickly becomes impractical when working with the complex, interconnected datasets that characterize modern scientific research.

For AI in labs to become truly useful, it must be able to access scientific data where it already lives: within the digital systems that manage experiments, samples, instruments, and research workflows.

The value of connecting AI to scientific data

A major shift is now taking place in how AI systems interact with software platforms. Instead of functioning as isolated chat interfaces, modern AI models are increasingly being equipped with capabilities that allow them to retrieve information from external systems and interact with other digital services. This evolution is what makes the next transformative phase of AI in labs possible.

One of the most important developments enabling this shift is the emergence of the Model Context Protocol (MCP). MCP acts as a communication layer that allows AI agents to securely connect with external platforms, retrieve information, and perform specific tasks within defined boundaries.  

For scientific organizations, this opens up entirely new possibilities. Rather than manually exporting experimental data or uploading documents into AI tools, researchers can interact with their research environment through AI agents that are able to access relevant information directly from the systems where it is stored.

How this works within the SMP

Bringing the value of MCP models into scientific workflows requires more than simply adding an AI interface. It depends on having a platform architecture that allows external systems to interact with scientific data in a structured and controlled way.

SciSure’s SMP was designed with this kind of interoperability in mind. The platform follows an API-first architecture, meaning that the functionality users see in the interface is driven by the same application programming interfaces (APIs) that external systems use to interact with the platform.

Since these APIs already expose structured access to experiments, samples, protocols, and other research records, MCP can act as a bridge between AI models and the underlying scientific data environment. AI agents can query records, retrieve protocol details and contextual information, and interact with the platform through clearly defined interfaces without compromising the integrity of the research data.

The result is a fundamentally different model for AI in labs. Instead of asking scientists to move data into AI tools, the AI can interact with scientific data directly where it already lives.

This ability to connect AI to structured scientific data transforms AI from a general-purpose assistant into a powerful research companion, capable of helping scientists navigate complex datasets, identify patterns across experiments, and surface insights that might otherwise remain hidden.

Explore Astra Iris, SciSure’s AI Support Assistant

What AI unlocks when it’s connected to data

When AI agents can interact directly with scientific data and research systems, the way scientists explore and interpret their information begins to change.

Instead of manually searching through ELNs, reviewing protocol histories, or cross-referencing multiple software platforms, researchers can ask questions about their experiments and receive answers grounded in their own experimental records.

With AI connected to structured research data through MCP, scientists can begin to unlock new capabilities such as:

• Rapidly retrieving experimental details
  Scientists can quickly access information from past experiments: such as reagent concentrations, protocol steps, or experimental conditions, without manually searching through historical records.

• Investigating unexpected results
  If an experiment produces an unexpected outcome, AI agents can analyze the procedures used, compare results across multiple runs, and highlight potential factors that may have influenced the outcome.

• Identifying patterns across experiments
  By analyzing large sets of experimental data, AI can detect correlations or trends that may not be immediately visible, helping researchers generate new hypotheses or refine experimental designs.  

• Connecting experimental outcomes to operational data
  AI can also analyze information from connected instruments and lab infrastructure. For example, environmental sensor data might reveal that an incubator was operating outside its intended temperature range during a failed experiment; an issue that could otherwise take significant time to diagnose.  

• Supporting deeper research analysis
  With access to structured experimental data, AI systems can help scientists explore their results more comprehensively: analyzing datasets, suggesting new experimental approaches, or identifying potential explanations for observed results.

These are just some of the capabilities that illustrate how AI in labs can evolve beyond a simple information tool into a powerful research assistant, helping scientists navigate complex datasets, uncover hidden insights, and accelerate the process of scientific discovery.

Ensuring Safe and Controlled AI Integration

While the potential of AI in labs is significant, introducing intelligent agents into scientific environments requires careful governance. Research data is highly valuable, experimental workflows must remain traceable, and organizations need clear control over how AI systems interact with their information.

For this reason, SciSure’s approach to AI integration focuses on providing powerful capabilities while ensuring that organizations retain full oversight of how AI is used within their research environment.

Key principles guiding the platform’s AI implementation include:

• Opt-in AI capabilities
  AI features are never automatically enabled. Organizations can choose whether to activate AI functionality within their environment, ensuring that adoption aligns with internal policies and regulatory requirements.

• Granular permission controls
  Through MCP, organizations can define exactly what AI agents are allowed to do within the platform. For example, an AI system may be permitted to read experimental data while being restricted from modifying records or performing other actions.  

• Support for local AI models
  Some organizations prefer to keep their data entirely within internal infrastructure. SciSure’s architecture allows customers to connect locally hosted AI models through the same MCP interface, ensuring sensitive research data never leaves their environment.  

• Flexibility to use preferred AI providers
  Rather than locking customers into a single AI vendor, the platform allows organizations to connect different AI models through MCP, enabling them to choose the systems that best align with their security, compliance, or performance requirements.

• Ongoing security and compliance commitments
  SciSure continues to strengthen its governance framework around AI, building on existing security certifications and working toward additional standards that address responsible AI implementation.

By combining powerful AI capabilities with strong governance controls, SciSure’s approach ensures that organizations can explore the benefits of AI in labs while maintaining the transparency, security, and oversight that modern scientific environments demand.

A new foundation for AI in labs

The future of AI in labs will not be defined by increasingly powerful models alone. It will be shaped by how effectively those models can connect to the scientific environments where research actually happens.

By connecting AI agents directly to scientific data and research workflows, new integration approaches such as MCP open the door to a more natural way of interacting with experimental knowledge. Scientists can move beyond manually searching records and begin exploring their data through intelligent systems capable of retrieving context, identifying patterns, and supporting deeper investigation.

At SciSure, this philosophy is guiding the development of new AI capabilities within the Scientific Management Platform. By combining an API-driven architecture with MCP-based integration, the platform enables AI agents to interact with structured scientific data while maintaining the governance, security, and flexibility required in modern research environments.

This shift marks an important step in the evolution of AI in labs. As platforms continue to integrate AI capabilities responsibly and securely, researchers will gain powerful new ways to understand their data, accelerate discovery, and navigate the growing complexity of modern science.

Interested in exploring how AI can interact directly with your lab’s scientific data? Get in touch with the SciSure team to learn how connected digital lab platforms are enabling the next generation of AI in research.

References

​​1. Lu, C. et al. Towards end-to-end automation of AI research. Nature 2026 651:8107 651, 914–919 (2026).

​2. Zhang, K. et al. Artificial intelligence in drug development. Nature Medicine 2025 31:1 31, 45–59 (2025).

​ ​

‍

Introduction

Analytica 2026 brought together over 30,000 visitors across four days in Munich, covering everything from analytical instrumentation to lab digitalization and quality infrastructure. It remains one of the few events where the full laboratory ecosystem is represented in a single venue.

This year, SciSure shared a booth with DataChaperone in Pavilion B2, right next to the Live Lab show. The setup reflected something we believe in: that structured data and connected workflows are stronger when they operate side by side. Here's a look at the week.

‍

Why being here matters

Analytica offers a rare overview of the entire lab landscape. From analytical methods and biotechnology to lab infrastructure and quality control, everything is represented in one place.

For us, being here is about staying close to real-world lab environments. It is where we hear directly from scientists and lab teams about their daily challenges, priorities, and expectations.

When asked about the impact of events like Analytica on our research community, Julia Wilke, Customer Success Specialist at SciSure, shared:

"These types of events strengthen our relationships with existing customers by giving them a space to exchange insights, explore new solutions, and accelerate the adoption of technologies that directly support their research goals. As for the rest of the community, it provides an excellent opportunity to see the value of these innovations in action and envision how they can benefit their own projects".  

How we support German-speaking labs and scientists

SciSure helps labs work in a more structured and efficient way by improving how data is captured, tracked, and managed, reducing administrative effort, supporting compliance, and making information easier to access and trust. We support German-speaking labs through a dedicated local team and by offering our platform in German, ensuring a smooth experience and clear communication. By providing flexible hosting options, including cloud, private cloud, and on-premises, we also enable labs to meet the strict data security and regulatory requirements of the German market.

DataChaperone strengthens this by focusing on data structure and governance. It's a natural fit for CROs, CDMOs, and analytical labs, or any organization where consistency in data handling is non-negotiable. It ensures that data remains traceable and usable across systems and teams, no matter how complex the workflow. Together, this creates a more reliable foundation for research and decision-making.

‍

What we saw and heard

The conversations throughout the week were focused and practical. Many labs are facing similar challenges:

• Increasing volumes of data

• Stricter regulatory requirements

• Disconnected systems that do not communicate well

• The need to scale without compromising data quality

There is a noticeable shift toward solutions that fit into a broader ecosystem rather than solving isolated problems.  ‍

Closing thoughts

Analytica remains an important checkpoint for the lab community. It highlights both the progress being made and the challenges that still need attention.

For SciSure and DataChaperone, it reinforces the importance of building solutions that fit within a larger ecosystem. Labs rely on systems that work well together, and that is exactly where we focus.

Next up, we'll be at Labs Live in London and Future Labs Live in Basel. If you're attending either event, come find us. We'd love to continue the conversation.

‍

‍

In most research organizations, materials are tracked, but they are not understood.

Traditional sample management and inventory systems record where materials are stored, but rarely capture the full intelligence those materials contain. Samples sit in freezers. Chemicals live in inventory systems. Animal models are logged in separate platforms. Biobank data exists somewhere else. Instrument outputs are stored in shared drives or isolated databases.

Everything is documented, yet nothing is unified. The result is fragmented insight, hidden risk, and missed opportunity.

Research Material Intelligence (RMI) is the discipline of transforming all research materials, from biospecimens and biobank collections to chemical inventories and animal colonies, into structured, interoperable, and analyzable data assets.

Instead of treating samples as inventory records, RMI connects lab sample management, biobank operations, and inventory workflows into a unified material intelligence framework.

It is not simply about knowing what you have. It is about unlocking what your materials know.

‍

Why lab sample management becomes fragmented in large organizations

Large healthcare and research institutions rarely run on a single, unified platform.

Biobanks operate independently from animal research facilities. Chemical inventories rarely integrate cleanly with experimental systems. Instrument data often exists outside the structured context of the materials it measures. Clinical systems and research systems speak different languages.

As a result, sample tracking, biobank operations, and inventory management often evolve separately rather than as part of a connected research infrastructure.

Each platform captures part of the story. None capture the full lifecycle of a material.

This fragmentation produces:

• Inconsistent identifiers across departments and systems
• Conflicting metadata between platforms
• Manual reconciliation that consumes researcher time
• Limited traceability from intake to outcome
• Low cross-domain visibility for leadership and operations teams

Many institutions today suffer from data abundance but intelligence scarcity. RMI exists to close that gap.

For organizations managing fragmentation across research and safety workflows, the challenge is often structural. A Scientific Management Platform that connects sample management, ELN, and EHS into a shared data layer can provide the foundation RMI requires.

‍

Pillar 1. How to standardize research material data across systems

You cannot build intelligence on top of inconsistency.

In most environments, material data is heavily dependent on local conventions. Naming standards vary by department. Metadata is incomplete or free-text. Identifiers change between systems. Over time, this creates invisible technical debt that makes integration fragile and analytics unreliable.

The first pillar of RMI is standardization.

For lab sample management to function across complex research environments, every material, whether a biospecimen, compound, plasmid, cell line, or animal cohort, requires:

• A persistent, unique identifier that follows the material across systems
• A defined metadata schema that enforces consistency at the point of entry
• Controlled vocabularies that eliminate ambiguity across departments
• Clear data ownership that assigns accountability for data quality

This is not administrative overhead. It is structural integrity.

Without standardization, interoperability becomes temporary. With it, material data becomes durable across every system it touches, from LIMS to biobank platforms to instrument databases.

Key takeaway: Standardized material data is the foundation of Research Material Intelligence. Without it, intelligence collapses into inconsistency.

Pillar 2. Why structured, relationship-aware data matters for research materials

Once standardized, material data must become machine-readable and relationship-aware.

The second pillar of RMI is structure.

A structured ecosystem does more than record attributes. It models relationships.

• A genetic construct links to a cell line.
• That cell line links to an animal model.
• The animal model links to experimental outcomes.
• Those outcomes link to clinical programs.

When these relationships are explicit, traceability becomes insight.

Too often, organizations store results separately from the materials that produced them. Instrument files are uploaded manually. Metadata is detached from its source. Experimental context lives in narrative form rather than structured fields.

Effective sample and inventory management requires that material data be structured at the point of creation, not reconstructed later. This is where tools like an electronic lab notebook become critical: they capture experimental context alongside material records in real time, preserving the relationships that matter for downstream analysis.

Key takeaway: Structured, relationship-aware data transforms materials from inventory entries into knowledge nodes.

Pillar 3. How to connect lab systems without rebuilding your tech stack

Most institutions cannot, and should not, rebuild their entire tech stack.

They operate with legacy LIMS platforms, animal management systems, ERP tools, clinical databases, instrument software, and data lakes. The objective is not forced consolidation. It is orchestration.

The third pillar of RMI is interoperability.

A healthy sample management architecture ensures that material data moves seamlessly across systems. This includes environments where biobank platforms, inventory tools, and experimental systems must all remain synchronized.

A connected ecosystem ensures:

• Systems communicate via APIs and SDKs
• Events propagate across platforms in near real time
• Material lifecycle changes trigger downstream updates automatically
• Data flows without manual reconciliation

When a biospecimen is accessioned, that event should not stop at the biobank. When an animal cohort changes, related systems should reflect it. When compounds are depleted, procurement should respond.

This is how data ecosystems become healthy.

Key takeaway: Research Material Intelligence requires a connected architecture where material data flows cleanly across systems, instruments, and databases.

‍

Pillar 4. How instrument integration strengthens material intelligence

Instruments generate some of the most valuable data in research environments. Yet, they are often treated as endpoints.

Data remains local. File formats are proprietary. Metadata about the material being measured is stored elsewhere. Context is lost.

The fourth pillar of RMI is instrument integration.

True RMI connects instrument output directly to structured material records at the time of experimentation. It preserves lineage automatically. It captures metadata programmatically. It eliminates manual uploads wherever possible.

When instrument data integrates seamlessly:

• Material records become dynamic, enriched with every measurement
• Context accumulates over time rather than decaying
• Reproducibility improves because experimental conditions are captured at the source
• Analytics become trustworthy because they rest on complete, connected data

Key takeaway: In RMI, instruments are not peripheral tools. They are core contributors to the material knowledge graph.

Pillar 5. How to sustain Research Material Intelligence with governance

Technology alone does not sustain RMI. Without governance, entropy returns.

Data models drift. Metadata becomes optional. Integrations quietly degrade. Fragmentation re-emerges across every system that touches research materials.

The fifth pillar of RMI is governance.

Organizations that succeed treat data governance as infrastructure. They establish oversight, protect schemas, monitor quality, and assign accountability for how research materials are defined and managed. In regulated environments, this governance must align with frameworks like 21 CFR Part 11 for electronic records, ISO 27001 for information security, and GLP/GxP requirements for data integrity across the material lifecycle.

Governance is not bureaucracy. It is durability.

Key takeaway: Sustainable RMI requires governance that protects long-term interoperability, data integrity, and consistency across every system in the research ecosystem.

Cross-domain visibility: the strategic payoff of RMI

When all five pillars of RMI are in place (standardization, structure, orchestration, instrument integration, and governance), cross-domain visibility becomes possible.

Instead of existing as isolated records across disconnected platforms, research materials become part of a connected data ecosystem.

• Biobank data connects to clinical outcomes.
• Animal model data connects to molecular profiling.
• Compound libraries connect to trial progression.
• Storage conditions connect to degradation analytics.

At this stage, materials become strategic assets rather than operational burdens.

Organizations can answer higher-order questions:

• Which sample attributes correlate with success?
• Which models predict outcomes most reliably?
• Where are material bottlenecks slowing discovery?
• Which compounds demonstrate cross-study reproducibility?

This is institutional learning at scale. Food Brewer AG, a cultivated food startup, achieved a 60% increase in R&D productivity after implementing connected sample tracking with lineage tracing and SDK-driven automation, turning fragmented material records into a unified system that compounds insight with every experiment.

Key takeaway: RMI transforms lab sample management and inventory from a tracking function into a compounding intelligence system for scientific discovery.

Data flow: the operational engine that powers RMI

You can standardize. You can structure. You can integrate. But if data does not move cleanly, intelligence does not emerge.

Data flow is the operational heartbeat of RMI.

In many institutions, data technically "exists" across systems, but it moves through exports, uploads, email attachments, or manual reconciliation. That is not flow. That is friction. This fragmentation is especially visible in sample management workflows, where material records, experimental data, and inventory systems often operate independently.

True RMI requires intentional, healthy data movement across the full lifecycle of research materials:

• From intake to storage
• From storage to experiment
• From experiment to analysis
• From analysis to decision-making
• From decision-making back into operational systems

When a biospecimen is accessioned, its metadata should propagate automatically across biobank and experiment tracking systems. When an instrument generates results, that output should attach directly to the originating material record. When a compound is consumed, inventory and procurement systems should update in near real time. When an animal cohort is modified, downstream analytics should reflect the change.

This is not about speed alone. It is about continuity and integrity.

Healthy data flow means:

• Events trigger downstream updates automatically
• Identifiers remain persistent across systems
• Context is never lost during transfer
• Metadata travels with the material
• Manual intervention is minimized

In RMI, data flow is not an integration afterthought. It is a design principle.

Without healthy data flow, standardization stagnates. Without healthy data flow, structure becomes isolated. Without healthy data flow, orchestration becomes theoretical.

With it, sample management and inventory systems become dynamic and continuously enriched.

Key takeaway: RMI requires intentional, event-driven, and interoperable data flow across systems, instruments, and repositories. When data moves cleanly and context travels with it, research materials evolve from static records into continuously enriched intelligence assets.

‍

The RMI framework: five pillars for turning materials into intelligence

Research Material Intelligence is a framework built on five core principles:

1. Standardization. Durable identifiers and controlled metadata.
2. Structure. Relationship-aware, machine-readable data.
3. Orchestration. Interoperable, event-driven systems.
4. Instrument Integration. Automated linkage of experimental output.
5. Governance. Long-term data integrity and stewardship.

Together, these pillars shift organizations from tracking materials to learning from them. When all five are in place, the result is cross-domain visibility: materials become strategic assets, and organizations can answer the higher-order questions that accelerate discovery.

[IMAGE: Combined RMI Framework + Data Flow diagram]

For a deeper look at how to build the business case for this kind of connected research infrastructure, the whitepaper Secure Leadership Support for a Unified Digital Lab Platform walks through ROI justification, stakeholder alignment, and phased rollout strategies.

From fragmented systems to Research Material Intelligence

Tracking materials is operational. Sample management and inventory systems help organizations store, catalog, and locate research materials. But Research Material Intelligence is strategic.

In an era of AI-driven discovery and increasingly complex experimental ecosystems, materials are not just physical assets. They are data anchors. Every sample, compound, and model represents a node in a growing knowledge graph that connects experiments, instruments, and outcomes.

Organizations that embrace RMI move beyond traditional tracking and build systems where insight compounds over time. Material data becomes structured, connected, and continuously enriched as research progresses. The approach is already producing measurable results: Arctic Therapeutics, an ISO 15189 certified biotech, centralized sample, inventory, and equipment management into a single system, saving two hours per week per researcher while strengthening clinical research quality.

Those that do not will continue reconciling spreadsheets, and wondering why discovery feels slower than it should.

The future of research will not be defined by how well organizations track their research materials, but by how effectively they transform them into intelligence.

‍