For too long, sustainability work has been consumed by data management. Collecting it. Cleaning it. Correcting it. Aggregating it. Analyzing it. Reporting on it. It's a treadmill that limits time for strategic work.
Data is supposed to serve sustainability leaders, not the other way around. Today, we’re flipping that script with the announcement of Watershed agents: a set of new AI capabilities that tackle the work that slows sustainability teams down the most. Our first release: AI that transforms messy, real-world data into clean, decision-ready insights. You can request a demo of these new agents and the rest of Watershed’s sustainability AI platform or join us on May 12 for a new AI Academy featuring sustainability-specific use cases to kickstart your strategy.
Because Watershed agents are purpose-built for sustainability, with domain expertise and business context baked in, they do more than save sustainability teams time. They also unlock a new era of impact by delivering strategic insights that were previously invisible, or accessible only with weeks of dedicated analysis.
“[The agents are] helping me save about 12 weeks per year,” said Christian Boothby, sustainability manager at Smiths Group. “That means I’ve now got time to work with our manufacturing and engineering teams on decarbonization strategy and energy efficiency projects across our sites.”
Video unavailable: Vimeo tracking consent required
The right technology is important—but it’s only part of the story. In a recent survey of 200+ global sustainability leaders, 37% of respondents highlighted an internal skills gap as a limiting factor in their adoption of AI for sustainability. That’s why we’re also announcing the Watershed AI Fellowship, a new accelerator program for sustainability professionals to become AI leaders in their organization and the broader sustainability community.
Watershed agents for data cleaning and analysis
Over the last year, we’ve launched AI capabilities across the Watershed platform, including Product Footprints for smarter product design and procurement decisions and end-to-end AI reporting for any ESG need. In December, we introduced an AI PDF scanner that can extract and validate data from bills and cut data ingestion time by up to 90%. Today, we're expanding our AI layer to cover two of the most time-consuming parts of the data management workflow: data cleaning and analysis.
Data cleaning. With Watershed agents, you can drag in messy data and get it cleaned, transformed, and measurement-ready in one session. The agents handle unit conversions, date formats, country codes, duplicates, missing values, and broken formulas. They de-duplicate across sources, reconcile overlapping datasets, and fill gaps with documented assumptions. Across our test customers, they've cut time to actionable data by 80%. One company completed a five-hour cleaning project in 20 minutes. And because they’re built with full visibility and controls, you can review every calculation and share with confidence.
Data analysis. In addition to giving you fast answers at your fingertips, our data analysis agents help you surface insights that can inform a better sustainability strategy. Using natural language, you can ask questions about your footprint, measurement methodology, or underlying data. Get year-over-year comparisons, identify decarbonization hotspots, flag anomalies, and more. AI answers come with clickable drilldown link or step-by-step reasoning back to the underlying data that informed it. One early customer told us that the data analysis agents gave them an insight that would have otherwise taken “three analysts and multiple weeks” to produce.
Emma Bayliss-Chan leads climate strategy at Royal Mail, one of the UK's largest fleets, with roughly 40,000 vehicles. Refrigerant data used to take her team two hours a month to process because it arrived in PDFs that couldn’t easily be ingested. Now, it takes moments. When last-minute business travel data arrived in an unfamiliar format the day before end-of-year reporting, the team thought it would take all night to decipher. The agents mapped the airport codes and built the file in minutes.
Emma said, “We can simply speak to [the agents] in plain language and say: Hey, I noticed our emissions went up this quarter; could you take a closer look into what the key drivers were? And they'll come back to us with information a lot faster than it would have taken to manually sift through all the data.”
Christian, at Smiths Group, oversees data from more than 250 sites globally. Consolidating Scope 1 and 2 data for monthly executive reporting, running gap fills, drafting commentary: that cycle used to take him a week. With the new Watershed agents, it takes a day or less.
Video unavailable: Vimeo tracking consent required
Watershed agents are purpose-built for sustainability AI
General-purpose AI can speed up almost any task, but sustainability work requires a higher level of rigor, specificity, and auditability. Watershed agents are built on the most comprehensive sustainability databases on the market, with over 500,000 emissions factors covering 95% of global GDP, and they were developed by a hybrid team of AI engineers and career climate scientists who know emissions methodologies from the inside.
Built upon Watershed’s specialized sustainability AI platform, these agents operate with a deep understanding of business and industry contexts, using the most current emissions data, scientific methodologies, and regulatory requirements. Knowing how critical auditability is to this work, every transformation carries full data lineage, changelogs, and system-wide hallucination checks, so the output is documented, defensible, and audit-ready.
As Emma at Royal Mail put it: "Watershed agents have our data context and the sustainability insights and intelligence that general-purpose AI doesn't."
Announcing the Watershed AI Fellowship
In addition to freeing sustainability teams to focus on more impactful work, we also want to ensure these teams have the support and training they need to become leaders in sustainability AI. To that end, we’ve opened applications for a new program: The Watershed AI Fellowship. Fellows will work alongside Watershed’s AI product leaders and a small peer cohort to perfect AI workflows for their sustainability goals. These Fellows will also help shape the next stage of sustainability AI, pioneering new use cases and setting standards for responsible use.
The fellowship is an eight-week program running May 12 through June 30, 2026. Throughout the program, fellows get small-group briefings with our climate science and policy experts, early access to new features, and direct feedback channels with our product team. By keeping the cohorts small, we hope to give participants a community of peers for knowledge-sharing long-after the program ends.
Applications are open now. The first cohort will be announced in May. If you are not selected for the first cohort, you will be considered for future cohorts.
At this point, the AI Fellowship program is open only to Watershed customers. If you’re not a Watershed customer, join our virtual AI Academy on May 12, where we’ll cover 6 sustainability AI use cases in 60 minutes.
Readying for a transformational decade
AI can be a force multiplier for sustainability, accelerating the work that has held teams back and making breakthroughs possible. We’re at a critical juncture. The technology has arrived, but it needs to be shaped and deployed by leaders who understand sustainability. We’re excited to see so many of our customers stepping into that role.
You can learn more about Watershed agents here and apply for the AI Fellowship here.
