From ELNs to AI Laboratory Notebooks

When our CEO said that “AI will eat the ELN,” he was pointing to a fundamental shift that’s already underway. For years, ELNs have been primarily passive tools. They were built to capture what happened in the lab, not to help scientists decide what to do next. That often meant researchers had to work around the software instead of with it.

Third-generation ELNs—what we call AILNs—change that. These are ELNs designed from the ground up to actively assist the scientist. They bring in scientific context, natural language interfaces, and embedded reasoning. The goal isn’t to replace scientists but to support them with faster answers and fewer clicks.

Sapio recently launched the first third-generation ELN, Sapio ELaiN, that allows scientists to interact with their notebook using natural language. They don’t need to know how the software works. They just ask for what they need.

That makes the notebook more than a record—it becomes a research assistant.

ELaiN is the first example of this new category, designed to reduce admin, surface insights, and support scientists as they work.

AI isn’t automating science end-to-end, but it is making it easier to ask questions, test ideas, and explore what’s possible. In many labs, even simple ideas require days of setup, formatting, and waiting for others to run models. That slows innovation and forces teams to prioritize only the most validated, lowest-risk hypotheses.

AILNs like Sapio ELaiN help reverse that trend. By using natural language to interface with scientific models and tools, scientists can ask “what if” questions and explore directions that would previously have felt out of reach. For example, they can use prompts to find similar experiments, run retrosynthesis using integrated models, or generate protocol drafts from an SOP, all without leaving the notebook or switching platforms.

This reduces the effort required to try something new. It doesn’t remove scientific judgment, but it lowers the barrier to exploration and helps reopen the space between inspiration and validation.

Sapio ELaiN improves efficiency by making the everyday tasks in scientific research faster and more intuitive, especially those that traditionally required multiple tools, platforms, or people.

For example, a scientist can use a prompt to turn a long SOP into a structured experiment—no more clicking through dozens of templates. What used to take 100 clicks can now take 1 or 2 prompts. The same goes for data search, protocol generation, and visualization.

ELaiN also integrates with commercial, open-source, and proprietary scientific tools. If a scientist wants to run retrosynthesis, perform codon optimization, or dock a structure, they can do it all from within the notebook. ELaiN connects to the right models, but the scientist stays in control, choosing what to run, when, and how to interpret the results.

None of this replaces scientific work. But it makes it faster, less manual, and more aligned with how scientists think.

Any AILN must meet the same standards as a traditional ELN for compliance, version control, and auditability. ELaiN was built with that in mind.

Every AI-assisted step—whether it’s an experiment design, a result interpretation, or a recommendation—comes with full provenance. That includes scientific suggestions (e.g., a retrosynthesis) that need to be generated by deterministic and well-tested scientific tools, keeping scientists in the loop, and ensuring they can always review, edit, and approve before moving forward.

This is especially important in regulated environments. ELaiN automatically builds an audit trail that aligns with frameworks like GxP and HIPAA (Health Insurance Portability and Accountability Act of 1996). Nothing leaves the platform unless explicitly allowed, and no result is final until a human says so. That’s how we combine AI acceleration with scientific and regulatory integrity.

The biggest challenge isn’t technical; it’s human. Scientists are often wary of new systems, especially if they’ve been burned by clunky ELNs in the past. The interface is unfamiliar, the workflows feel rigid, and the training takes time.

Sapio ELaiN helps solve that by removing the friction. You don’t need to memorize how the software works. You just describe what you want to do. That means scientists can be productive from day one, without weeks of onboarding.

Switching to a new ELN can be disruptive. But ELaiN’s natural language interface, built-in scientific tools, and prompt-driven workflows reduce the learning curve. Users can be productive from day one.

AI is progressing at such a speed that the prediction of future timelines is nearly impossible. Even two years ago, the progress in LLMs would have been hard to foresee.

What I do expect in the short term is that the AILN category will evolve from intelligent assistance to intelligent orchestration. Today, tools like Sapio ELaiN are helping scientists execute individual tasks more efficiently, generate experiments, retrieve data, and interface with models through natural language.

But as AI systems become more connected and more context-aware, AILNs will begin to support larger workflows across teams, tools, and datasets. Scientists won’t need to manually stitch steps together across software. They’ll be able to stay in the notebook and coordinate their work from one environment.

This doesn’t mean handing over control. The scientist will still define the question, guide the approach, and make the decisions. But the system will handle the steps in between—sourcing inputs, running models, formatting outputs, and flagging anomalies.

We expect that within three to five years, the gap between labs using AILNs and those that aren’t will be clear. Labs with AILNs will explore more hypotheses, iterate faster, and get to decisions sooner, while others continue to wrestle with disconnected tools and rigid workflows.

The introduction to this interview includes text that has been created with the assistance of generative AI and has undergone editorial review before publishing. Technology Networks' AI policy can be found here.