Waters sells AI peak integration as a paid upgrade to Empower users

Waters has launched AI-Powered Peak Integration for Empower, its chromatography data system, announced on 28 October 2026. It is a premium capability available only to customers on subscription.

What it does

Peak integration is the step where an analyst decides exactly where each peak in a chromatogram starts and ends. In regulated QC labs it is routine but slow, and much of it is still adjusted by hand. Waters says it is the largest source of variability in chromatographic results.

Get Software in Science every TuesdayThree-minute read. Free. One-click unsubscribe.

The new feature trains on each lab's own data to strip out noise peaks, merge high molecular weight peaks and edit baselines. Waters says the outputs are explainable and that models are version-controlled and do not drift until the lab chooses to retrain and approve a new one, which matters for audit trails.

The numbers

Waters is claiming up to a 95% cut in manual interventions. In a pilot of 400 chromatograms, interventions fell from 90 to 4. It also reports 76% greater accuracy, 50% faster review and 72% less variation between users. Empower has 500,000 users.

Enrolment for early access opens in December 2026, and validated, compliant deployment is expected in mid-2027.

"Our customers want the greater speed, accuracy, and productivity that AI brings, without giving up control of their data or the compliant audit trail behind it." Tina Wu, Senior Vice President, Waters Analytical Sciences

How it was built

Fortune reports that Waters spent around a year and a half working with an unnamed top-ten pharma company, with much of that time going on legal agreements and data safeguards before model training began. The same piece notes this is the second major AI feature for Empower, after AI/ML anomaly detection in February 2025, and that Waters launched AI Link, which brings Anthropic's Model Context Protocol to its informatics platform, two weeks earlier.

Why it matters

The interesting part is the business model. Rather than selling new seats, Waters is adding paid AI tiers to software that a very large base of regulated labs already runs. That is an attractive route for any incumbent informatics vendor, and it puts pressure on newer AI-first players to show they can match the compliance story, not just the model.

What it means for hiring

Premium AI tiers sold into an installed base make account management and expansion selling more important than new-logo hunting. We'd expect more demand for account managers and CS leaders who can run upsell motions in regulated labs.
The long data-sharing negotiation Fortune describes suggests that partnerships and legal capability with pharma QA teams is now part of shipping AI. People who can broker those agreements are in short supply.
Application scientists with chromatography and GxP experience who can explain validated AI to QA teams look like a strong hire for any vendor in this space.

Share Ask Jack about this →