EmergenCITY AI technology
Geospatial reasoning for proactive disaster planning.
EmergenCITY AI is Resilitix's flagship platform for Geospatial Resilience Intelligence. Ask plain-language questions about hazards, people, infrastructure, and essential services. ResilUS, the analyst agent inside the platform, connects those questions to spatial analysis and returns quantified, mapped findings your team can inspect and refine.
Spend less time assembling disconnected information and more time identifying where to focus before hazards become impacts.
How a question becomes an assessment
ResilUS plans the analysis and interprets the findings. Executed spatial queries supply the quantities. Each result can guide the next question, so an assessment develops as your understanding does.
Plan
Resolve the place, time period, entities, and criteria. Select relevant evidence and a supported sequence of analyses.
Compute
Run spatial queries to select locations, join compatible evidence, quantify exposure, and compare conditions.
Interpret
Explain what the findings support, how the evidence fits the question, and which conditions remain unresolved.
Refine
Use intermediate findings to investigate a priority area, examine facilities, compare locations, or adjust the assessment's scope.
- Math
- Computed quantities that show the scale of the finding.
- Map
- Geographic context showing where the selected conditions coincide.
- Memo
- A plain-language explanation your team can review and share.
- Audit
- The sources, time context, methods, and assumptions behind the assessment.
Follow a planning question through domain context, data selection, spatial computation, review, and refinement. Illustrative workflow; facility status remains subject to local verification.
Open full-size diagram (new tab)Read the diagram description
A planning question asks where wind exposure and outage vulnerability warrant closer assessment. Domain context supplies schema, disaster ontology, spatial joins, units, and valid time. ResilUS plans, executes supported operations, reviews returned evidence, and refines the assessment. The diagram connects forecast hazards, population exposure, building inventory, power vulnerability, and critical facilities with curated spatial data, including H3 and native points and networks. Spatial SQL performs filter, join, aggregate, and compare operations. The four outputs are Math, Map, Memo, and Audit. Dashed paths show requests; solid paths show results. Facility status must be verified locally.
The foundation behind a useful answer
EmergenCITY AI brings maintained spatial data, disaster-domain definitions, and reusable analytical procedures into one assessment workflow. Their connections make it possible to move from a planning question to an explicit calculation and a finding with a clear meaning.
Data organized to work together
The catalog documents 130+ geospatial data products across six families, covering hazards, people, critical services, infrastructure, geography, and native facility and network layers. A shared hexagonal grid (H3) and ZIP, county, and state crosswalks support analysis across these sources. Native points and lines retain facility and network detail.
Product availability, geographic coverage, and update timing vary. The count describes cataloged analytical products, including related representations, rather than independent upstream datasets.
Disaster concepts connected to calculations
An ontology supplies structured definitions and relationships for hazards, exposure, vulnerability, lifelines, and capacity. Schema mappings connect those concepts to data fields, units, time windows, spatial joins, and counting rules. This helps the analysis preserve the meaning of the question as it becomes executable.
Reasoning procedures that support follow-up
Reusable, versioned procedures guide evidence selection, spatial and temporal alignment, exposure analysis, and vulnerability comparison. ResilUS sequences these procedures and uses returned findings to determine which additional analysis would help refine the assessment.
For example: which facilities warrant closer review?
An assessment must specify the facility types, hazard period, exposure criterion, and counting unit. A facility in a partly inundated grid cell may warrant investigation; confirming flooding at that facility requires evidence at the relevant location and resolution.
What the evidence can support
Before a threat
Investigate baseline exposure, vulnerability, access to essential services, and preparedness gaps across the places your team is responsible for.
As a threat approaches
Connect supported forecasts with people, assets, and services to examine potential exposure and identify places needing closer assessment.
After impact
Compare available observed signals with baseline conditions to focus verification and subsequent planning. The documented after-impact example combines platform screening with external reports and analyst review.
Current product workflows and evidence of use are strongest in preparedness and anticipatory assessment. Internal operational tests and practitioner feedback inform continued refinement of the workflows.
Where EmergenCITY AI fits
Language models, connected tools, and geospatial capabilities contribute different parts of an assessment. EmergenCITY AI brings them together with disaster-specific data, definitions, reasoning procedures, and ResilUS orchestration.
Language, connected tools, geospatial analysis, and disaster context play complementary roles. These illustrative categories overlap.
Open full-size diagram (new tab)Read the diagram description
Language models explain, summarize, explore questions, and draft content. Tool-enabled AI retrieves information, uses connected tools, and assembles results. Geospatial AI analyzes spatial relationships, runs models and workflows, and produces maps and findings. EmergenCITY AI brings disaster data, ontology, and ResilUS assessment workflows together, with Math, Map, Memo, and Audit. Its disaster-assessment context spans hazard, exposure, vulnerability, and lifelines. These are illustrative categories with overlapping roles, not a vendor ranking.
See what supports the finding
Review the selected geography, data sources, relevant times, screening rules, calculations, and assumptions alongside the result. The Audit retains the analytical basis so your team can understand what was assessed and what remains to be checked.
Forecasts describe possible future conditions. Observations describe reported conditions. Baseline vulnerability indicates susceptibility. Keeping these meanings distinct helps your team interpret a finding without turning potential exposure into confirmed disruption. Missing coverage remains an information gap.
Assessments support screening and planning. Your team brings local knowledge, validates current conditions, and decides how the findings inform action.
Questions about EmergenCITY AI
What is ResilUS?
ResilUS is the analyst agent inside EmergenCITY AI. It interprets the planning question, selects evidence, sequences supported analyses, and synthesizes the findings. Resilitix is the company; EmergenCITY AI is the platform; ResilUS is its agent.
We already have hazard data and GIS. What does EmergenCITY AI add?
It connects maintained geospatial evidence, disaster-domain definitions, and reusable analysis procedures through a conversational workflow. Your team can investigate a planning question, inspect the analytical basis, and pursue a follow-up without assembling each assessment from the beginning.
How are the numerical findings produced?
Read-only spatial SQL executes the supported selections, joins, counts, aggregations, comparisons, and rankings. ResilUS interprets the returned evidence. A calculation can be reproduced when its query, data snapshot, and method are retained; refreshed evidence can change the result.
Can we review and share an assessment?
The browser application brings together Math, Map, Memo, and Audit, with interactive map and layer controls. PDF assessment export supports review and record-keeping. Findings should be read with their geography, time context, assumptions, and verification needs.
What do the animations demonstrate?
They illustrate the assessment workflow and the roles of different AI capabilities. They are conceptual diagrams, rather than measured disaster impacts or a performance comparison. The categories in the comparison diagram overlap.
Can EmergenCITY AI connect to another product or workflow?
That is a technical collaboration opportunity. The MCP tool gateway is in development; external interfaces, result formats, and deployment requirements are scoped with each partner. Start with one planning workflow and a joint evaluation of the contribution.
For technology partners
Bring disaster-planning intelligence into the tools your customers already use.
Explore how your models, data products, or user workflows could connect with EmergenCITY AI's maintained spatial foundation, disaster reasoning procedures, and assessment outputs. The objective is a geographically specific result with the evidence needed for review inside your users' workflow.
Available in the application
Supported planning assessments, interactive maps and layer controls, and PDF assessment export.
In development
The MCP tool gateway for partner workflows and a framework for evaluating geospatial assessment quality.
Scoped per collaboration
Bounded assessment interfaces, structured result formats, deployment requirements, and evaluation criteria.
Start with one workflow
Compare your existing workflow with the proposed integration. Evaluate analytical correctness, preservation of evidence, time to a usable assessment, expert intervention, and usefulness to the planner. Repeat in a second geography or scenario to examine reuse.