Data, methodology, and limitations
Coexist is built around transparent data, responsible analysis, and clear communication of uncertainty.
Data Sources
How Analysis Works
- 1The project area is defined
- 2Relevant environmental layers are identified
- 3Possible overlaps and conflicts are reviewed
- 4Evidence is summarized
- 5Confidence and limitations are shown
- 6Possible actions are suggested for expert review
How Analyze Uses AI
The Environmental Intelligence Overview on a Project Analysis result is generated in five steps. Understanding them is the fastest way to see what the AI can, and cannot, tell you.
- 1Environmental evidence comes from identified external, public data sources (OpenStreetMap, USFWS, USGS, GBIF, NY State DMV, NYSDOT FOIL, federal land records, and Coexist community reports), never from the AI itself.
- 2Coexist's own code performs the spatial calculations (distances, counts, screening bands) and structures that evidence into a consistent format.
- 3Deterministic, rule-based screening logic (not AI) turns that structured evidence into preliminary findings and considerations. This step is fully reproducible: the same evidence always produces the same findings.
- 4AI synthesizes those structured findings in the context of the specific proposed project (its type, planning stage, and description) and helps generate investigation questions and possible climate-adaptation relevance. The AI is only permitted to draw on the structured findings it is given; it does not have its own knowledge of the site.
- 5AI does not independently determine environmental compliance, permitting, regulatory applicability, or project approval, and cannot invent species, distances, regulations, or risk levels. Every number it references is checked against the real evidence before being shown; if that check fails, or the AI is unavailable, the overview is simply marked unavailable and the rest of the screening (all deterministic evidence and findings) still renders in full.
AI may help
- Summarize the structured findings for this specific project
- Prioritize which findings deserve attention given the project type and stage
- Suggest investigation questions grounded in the evidence
- Note possible climate-adaptation or ecological-resilience relevance, when supported by evidence
- Explain limitations already present in the evidence
AI does not
- Replace field surveys
- Replace professional judgment
- Determine regulatory applicability or provide legal approval
- Guarantee compliance or determine permitting requirements
- Invent species, distances, regulations, or risk levels
- Recommend specific mitigation measures
- Replace engineers, planners, or ecologists
Confidence
High
Multiple recent, verified data sources agree and cover the specific area closely.
Moderate
Some relevant data exists, but coverage, recency, or field validation is limited.
Low
Data is sparse, outdated, or largely illustrative for the specific area in question.
Limitations
- Incomplete datasets
- Uneven observation coverage
- Changing environmental conditions
- Seasonal variation
- Species-detection limitations
- Local conditions not visible in public data
- Prototype analysis
- Need for field validation
Scientific and Professional Disclaimer
Coexist provides preliminary decision support. It does not replace environmental assessments, field surveys, engineering analysis, legal review, regulatory processes, or professional judgment.