Environmental intelligence meets AI: A summary from a roundtable of EHS professionals

With new AI capabilities emerging rapidly, there can be pressure to find ways to deploy the technology. But the discussion emphasized that this can be the wrong place to start. 

Instead, innovation should begin with an outcome. The roundtable explored three broad categories: improving productivity, creating greater intelligence and enhancing the experience of employees accessing and using information.  

Applied to environmental management, that changes the question. 

Rather than asking, “Where can we use AI?”, an EHS leader might ask: “Where are our people spending too much time retrieving and reconciling environmental information?” Or, “Where could better visibility help us anticipate environmental or compliance issues rather than react to them?” 

Those questions can lead to more targeted use cases. The presentation identified examples including compliance reporting, facility and plant intelligence and knowledge assistants that make EHS and engineering information easier to access.  

For companies operating across multiple facilities or assets, this approach can also create an opportunity to connect information that has traditionally been managed separately and turn it into broader environmental intelligence.

The roundtable highlighted how the business case for AI can differ substantially across industrial sectors.

For oil and gas companies, the discussion focused on the pressure to accomplish more with leaner teams, as well as the potential to manage environmental information at the portfolio level. Participants discussed using consolidated data to examine root causes and track metrics such as greenhouse gas reductions.

For chemical and petrochemical companies, the conversation identified somewhat different priorities. Participants pointed to quality improvement, permitting and complex regulatory reviews as areas where AI can add value. One example discussed using AI to review a large regulatory permit application and identify deficiencies before submission. 

For manufacturing and other facility-based operations, the opportunity may begin closer to the plant floor. EHS data can span monitoring and sampling, incidents and spills, audits, corrective actions and facility management, creating potential applications for compliance reporting and facility intelligence. 

The use cases differ, but the underlying challenge is similar: valuable information exists across the organization, yet accessing, connecting and interpreting it efficiently can be difficult.

One of the most useful ways to think about AI's potential for EHS is through the progression of analytics maturity.

The discussion outlined four stages:
Descriptive: What happened?
Diagnostic: Why did it happen?
Predictive: What could happen?
Prescriptive: What should happen?

Much environmental management has traditionally depended on the first stages: collecting monitoring information, producing reports, demonstrating compliance and investigating issues after they occur.

Environmental intelligence creates the potential to push further. Participants discussed an operational example in which AI could move from understanding what happened and why, toward predicting an equipment need and eventually recommending when maintenance should occur to achieve a desired outcome. They also discussed fleet management, permitting requirements, asset assessments and AI-enabled compliance scheduling. 

The same progression provides a useful ambition for environmental programs: not simply reporting environmental performance, but using connected information to identify patterns, anticipate risks and inform the next decision.

Getting there depends heavily on the information underneath the technology.

Environmental programs generate significant structured data, but they also rely on information contained in documents, reports and other unstructured formats. During the roundtable, participants discussed the need to bring structured databases together with unstructured documentation to support more effective forecasting. 

This is particularly relevant to EHS because important information can reside across permits, compliance documentation, monitoring and sampling records, safety and incident information, spill and SPCC records, audits and corrective actions. 

The goal does not have to be a massive overhaul of every system. GHD's presentation advocated an iterative approach: define the outcome, prioritize the use cases, identify capability gaps, test through lean proofs of concept and scale from there. It also emphasized avoiding large, multi-year data transformations in favor of building unified data layers around priority use cases.

"Environmental intelligence does not require organizations to solve every data problem before they can create value. It requires them to identify which environmental decisions matter most and connect the information needed to improve them."

Roundtable participants identified security, cybersecurity, confidence in AI-generated outputs and the ability to check work as significant concerns. They also discussed the risk of “bot sitting,” where applying AI to the wrong problem creates so much human verification that the expected productivity benefits disappear.

For EHS teams, human oversight is especially important when outputs intersect with regulatory compliance, safety or operational integrity. GHD's presentation emphasized keeping humans in the loop and designing governance around intended outcomes, with consideration for safety, cybersecurity and regulatory compliance. 

AI therefore does not remove the need for environmental expertise. It can change where that expertise creates the greatest value.

When technology takes on more of the work involved in finding, organizing or interpreting information, professionals can spend more time applying context, interrogating results and making informed decisions. The goal is to give environmental professionals better intelligence with which to navigate increasingly complex operations.

For chemical, oil and gas, manufacturing and other industrial organizations, the value of AI may ultimately be less about the technology itself and more about the environmental intelligence it makes possible.

The roundtable showed that organizations are at different stages of that journey. Some are experimenting, others are moving applications beyond the pilot stage, and concerns around trust, culture, skills and governance remain.

But the path forward does not have to begin with a sweeping transformation. It can start with one meaningful environmental challenge: a manual compliance process, fragmented facility information, a forecasting need, difficult-to-access institutional knowledge or an opportunity to see environmental performance across an asset portfolio.

Start with the outcome. Connect the information. Keep people and environmental expertise at the center. Then determine where AI can make the process more intelligent.

That is where the opportunity moves beyond AI adoption and toward an EHS function that can turn more of its environmental data into environmental intelligence, and environmental intelligence into better decisions.

We offer interactive workshops for organizations looking to identify high-value AI opportunities and create a clear path from ideas to action. Through the workshop, teams align on the outcomes that matter most, develop and prioritize practical AI use cases, and select one or two high-value opportunities to advance through a focused 90-day action plan. Contact our team to explore where AI can create meaningful value for your organization and identify the right place to start.