As environmental due diligence continues to evolve, long-standing challenges around City Directory research for Phase I Environmental Site Assessments (ESA) are finally becoming more tractable through better data infrastructure and governance. At the same time, AI adoption across environmental workflows remains uneven, a trend highlighted in LightBox’s recent benchmark survey on AI use in the sector.
Ahead of the EBA conference, we asked Alan Agadoni to draw on his decades in the environmental due diligence business, and as a speaker at the conference, to share a practitioner’s perspective on how City Directory research has changed, where AI is delivering real value today, where it remains overhyped, and what excites him most about where the environmental due diligence and CRE sectors are headed.
Why was City Directory research for Phase I ESAs so difficult historically and what’s meaningfully changed?
Historically, it was difficult because the unstructured city directory data existed across many pages in static, analog books, or manual scans, while the analytical task required parcel-level certainty. Street name changes, address renumbering, shifting municipal boundaries, inconsistent coverage, and listing errors forced Environmental Professionals (EPs) to rely on manual cross-referencing and inference. The process was time-intensive, error-prone, and highly dependent on individual expertise, creating friction and defensibility challenges despite the high value of the information.
City Directories in LightBox Live (officially launching on February 16) have vastly improved not only the source data itself, but the infrastructure applied to it. OCR, curation, and geospatial normalization now convert directories into structured, searchable datasets. Addresses can be tracked spatially across time, machine learning helps identify commercial uses from residential, and directory data can more easily be cross-referenced with other historical sources. This shifts city directories from a difficult interpretive exercise into a repeatable, auditable, and workflow-integrated component of modern Phase I ESA research.
LightBox just completed the industry’s first benchmark survey on AI use for Environmental Due Diligence. What are some key takeaways from the findings, and where is AI currently helpful versus overhyped?
The responses were unsurprisingly polarized, but we’re not seeing much blind enthusiasm or wholesale rejection. About half of respondents are already using AI, but very deliberately. The use cases are narrow and practical; research, reviewing large document sets, drafting non-opinion-based report sections. In other words, AI helps EPs and risk managers move faster through information, not make judgments.
What’s also telling is the confidence gap. Only a tiny fraction of respondents view Chatbot LLMs as truly effective for due diligence today, while the overwhelming concern is accuracy and misinformation. That reinforces a core theme: environmental due diligence is still a defensibility exercise. If you can’t explain or stand behind an output, it’s not usable in a regulatory, transactional, or litigation context. Proper governance and training in responsible AI use is badly needed, but our industry, like most, is lagging behind.
Where AI is genuinely helpful right now is as a pattern recognition force multiplier: triaging portfolios, hunting through decades of voluminous records, and accelerating discovery of potential data gaps or hidden risks. That’s real value.
Where it’s overhyped is the idea that AI can replace expert judgment, interpret complex environmental conditions per ASTM standards, or independently determine recognized environmental conditions. The survey makes it clear: professionals want AI to support decisions, not make them—and they’re right to be cautious.
A full report on the survey findings will be published later this month.
What excites you most about where the CRE and Environmental Due Diligence sectors are headed?
I am fascinated by how the broader industry trends in AI, ESG, and data-driven workflows are converging with real-world technology developments to reshape both investment decisioning and environmental due diligence in practical ways. Commercial real estate is moving from siloed, manual processes toward integrated, intelligence-rich platforms that help teams act faster with more confidence. LightBox’s data and platform solutions are right in the middle of it: merging authoritative property, spatial, and environmental data for deeper market insights and more reliable underwriting.
At the same time, sustainability and regulatory pressures are elevating environmental risk into a core component of CRE due diligence. By integrating historical records, regulatory sources, and property physical data, LightBox Live helps accelerate Phase I assessments and better inform ESG-informed investment strategies.
Together with our structured appraisal extraction through Fundamentals, this means the CRE sector is doing more than digitizing workflows. Environmental and market intelligence are increasingly embedded across valuation, lending, and risk decisions, aligning CRE technology evolution more closely with capital allocation and risk management.
Editor’s note: City Directory research will be available in LightBox Live beginning February 16, introducing a mapped, reviewable workflow for Phase I ESA research.
