Analytical laboratories have made major gains in sample preparation, acquisition, and instrument automation. But for many teams, the biggest workflow constraints now appear after data collection.
For Aude Tartiere, Head of Commercial, Expressionist at Genedata, the pressure has shifted downstream, where scientists need to process, review, interpret, and report increasingly complex LC-MS data. “Instrument vendors have been doing a great job,” Tartiere remarks. “I don’t think the acquisition is the issue anymore. It’s definitely on the analysis side.”
That shift matters as laboratories adopt more automated and AI-enabled workflows. LC-MS methods can generate rich data across molecules, methods, and sample types, but laboratories still need connected systems that preserve context from registration through reporting.
Tartiere describes the goal as an end-to-end workflow that connects molecule information, metadata, sequence data, analytical files, processing results, and downstream systems. In practice, this can reduce manual steps and support more consistent data handling. “That’s where you really add value,” she notes. “You reduce the time for analysis and reduce the chance of making errors.”
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
Why Review Still Needs Human Expertise
Although automation continues to advance, Tartiere argues that many analytical workflows still need expert review. The challenge lies in reducing the time scientists spend checking results without removing scientific judgment from the process. “We still need a human to check,” she advises. “What we are working on right now is trying to embed the logic of an expert into the workflow.”
That approach can guide scientists through review steps, flag issues, and help users focus attention where it adds the most value. Instead of replacing the reviewer, the workflow can support faster, more consistent decisions.
Reporting creates another common bottleneck. Tartiere points to the historical burden of creating Excel files, copying results, pasting figures, and rebuilding reports for each workflow. Automated reporting can remove those manual steps when laboratories need consistent outputs in defined formats. “You now have the ability to create customized reports in exactly the format you need,” she asserts.
AI Needs Better Data Foundations
As laboratories explore AI, Tartiere urges a practical view of what the technology requires. AI depends on data quality, structure, and scale. Poorly curated data will limit what AI can achieve. “Data needs to be high quality, and sometimes that’s not the case,” she cautions, adding that if the data is not verified and cleaned properly, it simply won’t be useful.
For AI-ready workflows, Tartiere highlights three requirements: high-quality data, strong metadata, and scalable systems. LC-MS software must support data processing while preserving the context that makes results usable across teams and systems. “Data needs to be annotated well with good metadata,” she explains. “In AI, you need a lot of data, so you need a scalable, powerful system.”
That point cuts through some of the hype surrounding AI in analytical science. Tartiere sees value in AI, but only when teams apply it to defined problems. “AI is super powerful, but it’s not a magic weapon,” she warns. “You do need to solve a problem and add value.”
Without that focus, laboratories risk investing in projects that deliver little more than existing algorithms or established workflows. “We should leverage AI for the right reasons,” Tartiere adds. “But I think we should really embrace the journey of AI as well.”
From Faster Processing to Better Decisions
For Tartiere, meaningful innovation in mature mass spectrometry will come from better decision-making, especially as laboratories combine data from multiple analytical technologies. “The next step is about decision-making,” she comments. “That’s where we can really make an impact.”
Mass spectrometry has a central role to play because of its sensitivity, throughput, and analytical power. But Tartiere expects the next phase to involve broader integration across mass spectrometry, chromatography, and other laboratory technologies.
“It’s going to be about consolidating different technologies,” she notes. “That’s where AI makes sense, because it can make a decision faster using chromatography, mass spectrometry, and other techniques.”
For drug development teams, that could mean quicker decisions, fewer manual handoffs, and stronger data continuity across regulated workflows. It could also help laboratories reduce failure risk by improving how scientists capture, interpret, and reuse analytical data.
The opportunity does not start with AI alone. It starts with connected data infrastructure, consistent metadata, and workflows that bring expert logic closer to routine analysis.
As Tartiere frames it, the future of LC-MS innovation depends less on acquisition alone and more on how laboratories manage what happens after the run.
