20th Nov, 2026 Read time 8 minutes

When Your EHS Manager Leaves, Can AI Take Their Place?

By Daina Huntington, Manager of Solutions Consulting and Senior EHS Advisor, SciSure

 

One evening, browsing through LinkedIn, I came across a story that EHS managers know all too well: a colleague left, a new starter took their place, and the last person’s work had lived entirely in their personal inbox, laptop, and in their head.

What struck me was how many responses came from people who had been the replacement. One EHS leader shared that when they stepped into their role, corporate templates were available, but there was no clear way to track what had already been done or what still needed attention. As a result, they had to recreate processes just to piece together the history and understand where things stood.

For over a decade, I’ve worked with research institutions and biotech companies to strengthen lab safety programs, and improve chemical inventory management. I’ve seen firsthand how smoothly these transitions can go – and how challenging they can become when they don’t. Lately, the question I hear at safety and operations events has a new twist: if we lose our EHS manager, can AI fill the gap?

My short answer: AI can make transitions faster and easier, but it can’t fill in gaps when information was never documented.

Why the AI question keeps coming up

Over the past year, I’ve attended numerous safety and operations events, and AI is the most common topic. People are excited about its potential to save time, but they’re also cautious because they know it doesn’t always get things right. 

Opinions seem fairly split: people see the benefits of AI, while recognizing its limitations. At CSHEMA this summer, the first audience questions in the vendor AI sessions were about accuracy and reliability. Where does this output come from, and how can we trust it? That skepticism is healthy, especially when considering how AI can support knowledge transfer during staff turnover and transitions. 

What AI does well in a handover

Used effectively, AI can make it much easier for someone stepping into an existing EHS program. It can:

  • Search years of inspection reports, incident records and training data in seconds
  • Identify open issues, overdue or recurring actions for a specific lab, building or department
  • Draft a first-pass handover note from the records that exist
  • Capture details from a chemical container label using a phone photo, so intake doesn’t stall while a role sits vacant

For a new EHS manager facing hundreds of labs, that’s a real head start. This can save significant time and provide a clearer picture of where things stand. 

But AI is only as useful as the information it can access. If records are missing, outdated, or scattered across different systems, even the best AI tools will struggle to provide reliable answers.

What walks out the door

What leaves with a person is usually the decisions and reasoning behind the forms.

Take biosafety. In a research organization, an Institutional Biosafety Committee (IBC) reviews work with biological agents and recombinant DNA before it can start. A new biosafety officer shouldn’t have to reconstruct why the IBC approved a particular protocol, what restrictions it imposed, or whether earlier concerns were ever resolved. In a lot of organizations, that history sits in email threads, meeting notes, and one person’s memory. They need to know why it was approved, what conditions were required, and whether previous concerns were resolved.

Inspections work the same way. Maybe the report says a fume hood sash was left open.  But does it show that this was the third occurrence, that the PI committed to correcting it, or that follow-up happened through email? Without that history, a new EHS manager is left trying to piece together what happened and what still needs attention.

AI can’t summarize information that was never documented. If it only has half the story, it can only give you half the picture, even if the answer sounds complete.

Put the history where the work happens

The fix starts with where records live. When the decisions, actions, and follow-ups are documented together, a new employee understands the context without rebuilding it. It also gives AI more complete, reliable information to work with. 

In practice, that means:

  • Approvals carry their history.
    The questions reviewers asked, the conditions they set, and every amendment and renewal stay attached to the registration, providing a clear record of decisions.
  • Inspections carry their follow-up.
    Findings, photos, corrective actions, and related correspondence sit with the inspection, so the next inspector can see recurring challenges, from months & years past.
  • Records belong to the organization.
    Chemical inventories, training records, and documents are owned by the lab, department, or site. When someone leaves, their access ends and the records stay.

The goal is simple: Preserve institutional knowledge, make transitions easier, and give both people and AI the complete picture they need to make informed decisions.

Check what the AI is standing on

The other critical factor is accuracy.  In safety, incorrect information doesn’t just create inefficiencies. It can put people at risk.

An incorrect chemical hazard classification could lead to improper storage, incompatible chemicals being stored together, inadequate PPE, or employees unknowingly working with hazardous materials without the right controls in place. These aren’t simply data errors. They can have real consequences for researchers, employees, and the organization. 

Much of the hazard information AI tools draw on comes from public sources and safety data sheets (SDS). Those are a good starting point, and AI is quick at pulling them together, but that doesn’t mean every result is accurate or appropriate for the specific chemical or work being performed.

That is why human verification matters. 

An example of chemical inventory software is one that uses AI to capture information from a photo of a chemical container label. Before that information is saved, a person reviews and confirms it. The hazard data associated with chemical records also undergoes human review.

AI can accelerate the work, but it shouldn’t replace the judgment and accountability required to keep people safe.

The goal isn’t just faster access to information. It’s making sure the information people use to make safety decisions is accurate, reliable, and trustworthy.

The handover test: Would your EHS program pass?

If your EHS manager left tomorrow, could their replacement find last year’s investigations, outstanding corrective actions, and the reasoning behind important safety decisions without having to track someone or multiple documents down?

If you’re not sure, your organization may be more dependent on individual knowledge than you realize. 

Whether you’re assessing your current processes or evaluating EHS software vendors, here are five questions worth asking:

  • Can you trust your hazard data?
    Where does the hazard data come from, and who reviews it before employees rely on it?
  • Can you trace biosafety decisions?
    Can a new biosafety officer see why a protocol was approved, what conditions apply, and whether previous concerns were resolved, without digging through multiple solutions?
  • Can you track historical inspections through resolution?
    Are inspection findings, corrective actions, and follow-up correspondence stored with the inspection? Can the next inspector see repeat findings?
  • Does institutional knowledge stay with the organization?
    Who owns the records: the person or the organization? What happens to access when someone leaves? Do records remain accessible to authorized personnel while their system access is removed?
  • Is there human oversight of AI?
    When AI fills in a field or generates safety information, does someone verify before employees begin to rely on it?

The takeaway: AI can help new employees get up to speed, but only when the underlying information is complete, accurate, and accessible.

Before investing in more AI, make sure your organization is preserving the knowledge behind its safety decisions.

Because a strong safety program shouldn’t only depend on who happens to be in the role.

 


About the Author: Daina Huntington

Daina Huntington is Manager of Solutions Consulting and Senior EHS Advisor at SciSure, with more than a decade of experience helping leading research institutions and biotechnology organizations strengthen laboratory safety, regulatory compliance, and operational excellence. Her expertise spans chemical safety, biosafety, risk management, and technology-driven safety programs. She partners with EHS executives, R&D leaders, and laboratory operations teams to modernize safety practices, improve organizational oversight, and build more connected, sustainable safety programs.

Brands who we work with

Sign up to our newsletter
Keep up to date with all HSE news and thought leadership interviews