Anchoring down at NAEP 2026
I spent last week enjoying the fresh air, abundant sunlight, and the company of some fine folks in Anchorage at the National Association of Environmental Professionals conference. The conference brings together an array of attendees who care about environmental reviews: agency staff, consultants, technologists, and more. I came with lots of questions, got lots of answers, and had a great time even as my internal clock went haywire.
The half-empty toothpaste tube reality of NEPA
If it hadn't become clear to me years ago when interviewing NEPA professionals, this conference made it abundantly* clear: the environmental work that used to happen inside the formal NEPA process is increasingly happening before it starts. (*Don't take that as an endorsement, yo.)
Scott Phillips at SWCA described the paradox. For Department of the Interior energy emergency procedures, agencies want packages that are "shovel-ready." This means all environmental work is done before the application arrives. This front-loading helps reduce risk and is operationally successful, but the tools practitioners use to manage their work (project tracking systems, document management platforms, basic spreadsheets) are built almost entirely for the formal NEPA phase. Pre-NEPA work stays invisible to many of these systems.
Taylor Horne, Laurie Cummings and Ben White described Alaska DOT's experience with Planning and Environmental Linkages. A PEL is a federally authorized process that lets transportation agencies conduct scoping, alternative development, and community engagement before the formal NEPA clock starts. They shared three project examples: the Seward Highway corridor in Anchorage, where a community living with eight lanes of highway since the 1970s finally got a formal channel to say "we don't want this"; Shishmaref, a barrier island with no road access, where two years of pre-NEPA field work prevented the wrong alternative; and a Kodiak corridor, where the project changed from "replace two bridges" to "what does this road actually need to do?” In each case, the PEL let the project change before making commitments and starting the review period.
No matter how many times the Powers That Be work to shrink review timelines, they forget the theory of constraints that says when you limit a specific step, you accumulate excess capacity in the rest. So when you focus all the attention on constraining the duration of the environmental review process without actually reforming it, the result is that work gets moved OUTSIDE the official duration. You can squeeze a half-empty tube of toothpaste all you want and move the contents around, but nothing changes if you leave the cap on. [That might be a very real thing I did when I was a little sleep-deprived while dealing with the extended sunlight and my circadian rhythms were in full-on samba mode.]
AI’s here, and everyone’s still figuring it out
Environmental professionals are using AI, and the tools are still developing. Plenty of teams are using it to sift through large datasets, conduct early-stage research, and explore geospatial layers. But for every use case, there were plenty of questions about where the line is. For every tool or use case, there were questions about legal determinations, methodologies, and the role of professional judgment with the practitioner. (Spoiler alert: The professional should always have the last call.)
Meredith Parkin, J.D., PMP described her team's process: drafting preliminary project descriptions from engineering drawings and comparable EISs, synthesizing large datasets, checking consistency of terminology across documents, and categorizing public comment responses by theme. My information architect heart also appreciated her three-bucket taxonomy for AI tools (general LLMs, enterprise workflow AI, and specialized environmental tools). Mehmet Secilmis, MS, GISP brought the geospatial story. A company called Skytex used Esri's wetland identification model on Tennessee data and found 800,000 additional acres of predicted wetlands that the National Wetlands Inventory had missed. Tennessee accepted the results. The dataset tells field teams where to look, not what they'll find when they get there.
Other great sessions included Lauren Schramm talking about how AI slop is already seeping into the permitting process and the Pacific Northwest National Laboratory and the Idaho National Laboratory showing developing tools and datasets aimed at a range of challenges in the permitting space. (Full disclosure: As a part of EPIC, I am collaborating with PNNL on researching the challenges permitting professionals face with tools and data.)
The governance problem keeps showing up in different clothes
Eric Beightel gave a session on the current federal environment, explaining the pendulum swing of environmental policy. While the volume of executive orders in this administration's first year exceeded every prior president's four-year total since Clinton, the back-and-forth isn't new. He explained that Secretary Berger's order, which routes all renewable energy authorizations through the Secretary's office, creates a deliberate processing bottleneck, showing how policies become workflow decisions.
The volatility follows a structure. Administrations adjust three things: the scope of review, the required documentation, and the mandated coordination steps. They tilt those adjustments toward sectors they favor and away from those they don't. If you understand those three dials, the current environment is less chaotic than it looks.
Beightel's reform argument follows: durable decisions should be protected from reversal except by courts that find fault. Statutory reform is better than a statutory rollback. Oversight bodies should be empowered to compel decisions rather than just coordinate. That last point landed differently after hearing about his FPISC experience. The permitting council could post project schedules, convene agencies, and troubleshoot. It couldn't force a decision.
Wesley James Furlong, Senior Staff Attorney at the Native American Rights Fund, described what happens when that same design constraint shows up in law. After the D.C. Circuit ruled that CEQ's binding rulemaking authority rested on a Carter-era executive order rather than a statute, every major federal agency now writes its own NEPA regulations. The regulatory landscape varies by lead agency and remains in flux, leaving most groups in reactive mode.
Fun fact: If you're building permitting technology, this fragmentation is a design constraint to design for, not around.
A plug for NAEP
NEPA is complex, no matter how you try to explain it. Explaining permitting is one of those "how much time do you have" questions. A good story about why they are important beats a policy explanation every time. I love these stories, and a conference is one of the best places to hear them. The firsthand examples, good and bad, gave insights that don’t travel as well through blog posts and articles. Explaining how people are going to pull over to watch bears, no matter how you design a highway, brings non-standard scenarios to life in a very furry way. Someone explaining, in lurid detail, how many times a system rejected documents until a software developer was called in and realized a flag was set to always reject a field, no matter what, is better than any demo. Even just the shared space of “hands up, who has been told to ‘go faster’ with no explanation of how,” and the “can I get an amen!” is worth the admission price.
Thank you to everyone who shared their stories, connected me with others, and offered a beverage or two along the way. We’re all in this together. Here’s to you. Thanks to Lauren, Matt, and others for socializing; to Gerbil at the Blarney Stone for a great lunch-time story; and to the NEAP board for letting me watch their meeting. Lastly, an apology to Ted Boling for missing his talk, I hope to get a version in-person soon during a DC visit.
Note: If you were at NAEP and I didn't get a chance to say hi, say hello, or let's have a virtual coffee. Now, if you'll excuse me, I have to go research a new coat purchase.
