Dotmatics ships new scientific workflows in days to weeks by embedding Sigma in Luma

Dotmatics made its name on hyper-specialized scientific point solutions. The harder problem was unifying the disparate data those tools produce inside Luma, its lab orchestration platform on Databricks. A homegrown visualization layer didn't scale — so Dotmatics embedded Sigma instead. The payoff: configuration replaced code. Demos and client go-lives now ship in days to weeks, governance is baked into the config layer, and scientists explore vendor-neutral instrument data and commit decisions straight back into the platform.

The Challenge

Dotmatics is a company built by scientists for scientists. Over years it produced some of the best point solutions in the market for highly specialized scientific analysis, from flow cytometry to chromatography. As research organizations scaled with that hyper-specialized data, a strategic gap appeared: none of it could be connected. Dotmatics answered with Luma, a lab orchestration and instrument integration platform built on Databricks that helps scientists create and collate data across sources and lay the foundation for the AI work they want to do next.

Dotmatics' own expertise sat in scientific data management, modeling, and analysis, but a data platform does not feel real to a scientist until they can see their data in it. The team experimented with bespoke visualization and charting to show what data looked like as it moved from vendor-specific formats into something unified. It worked as a demonstration, but scaling challenges were obvious. Every change lived in the code base, which meant long release cycles and heavy maintenance overhead, and it pulled engineering attention away from the science and data problems Dotmatics knew best.

The Solution

Dotmatics needed a configuration-based layer that would plug into the architecture it already had and own the final mile of visualization and exploration. Sigma became that layer inside Luma: Dotmatics creates, prepares, and shapes the data, and Sigma is where customers see it, explore it, and act on it. Sigma's plugin architecture let the team build highly domain-specific extensions that no one outside scientific research would need, and extensive use of Actions turned those views into workflows that write back into Luma.

Instrument data shows what that looks like in practice. Labs receive the same underlying research in different vendor formats, sometimes as complex binary files. Dotmatics' parsing engine extracts each vendor format, then data models in Databricks and Luma harmonize them. Sigma presents the harmonized, vendor-neutral result on screen, so a scientist can examine the peaks from an instrument such as an X-ray diffractometer, select the data that matters, and commit that decision back to the database through plugins and Actions, kicking off downstream processes elsewhere in the platform. Exploration, decision, and commitment happen in one place rather than across disconnected systems.

Rolling this out reframes how customers think about scientific reporting. Like healthcare, science is accustomed to static reports, so Dotmatics starts each engagement with a scoping exercise built around a different question: not what does your current system do that you want rebuilt, but what does ideal look like, and how do you want to change the way you work and the science you do. Teams begin with something simple, grow into more complex workflows, and mature toward automated and AI-assisted decision-making.

The Results

Because delivery is configuration rather than code, Dotmatics pulls together demos and client go-lives in a matter of days to weeks, where changes previously required much longer cycles and heavier maintenance. That matches Luma's own ethos of a config-based platform with governance baked into the config layer. For customers, it means going live faster than competing vendors, and faster and at higher quality than an internally built tool. And because the pipeline is now automated from data generation through exploration to committed decisions, the decisions themselves are better.

The biggest thing we're seeing is higher quality decisions that are being made faster, with a lot lower overhead to implement and maintain.

— Michael Fritz, Principal Product Manager, Lab Orchestration, Dotmatics

Governance and configurability are what Dotmatics would miss most if the layer disappeared. Drag-and-drop workbook building democratizes delivery to people who are computational but not software engineers, which describes much of the life sciences workforce, without giving up the quality controls clients expect. The stakes justify the rigor: a bad decision in this domain can end with a suboptimal drug reaching a patient. Rebuilding that configurability, composability, and governance in-house would be possible, but it would pull focus from Dotmatics' core competencies and slow its ability to go to market.

Now part of Siemens, Dotmatics is being pulled from early research further downstream into development, manufacturing, and scale-up, and beyond life sciences into other scientific domains. Looking ahead, the team is watching Sigma's AI work closely: AI-assisted workbook building to accelerate the complex search, filtering, and integration workflows Dotmatics builds today and to extend go-lives past its services team to scientific end users, and natural-language exploration through Sigma Assistant to surface insights the sheer scale of scientific data would otherwise hide. On the engineering side, Workbooks as Code is the most anticipated: a higher-quality, lower-risk way to move app configurations across clients and tenants than the approach Dotmatics uses today.