Phase 2: Business Analysis (“Use Case Development”)

Objective

The objective of Phase 2 is to conduct a thorough analysis of the current business environment to identify and prioritize specific, high-value use cases for GenAI. This phase bridges the gap between the high-level business problem defined in Phase 1 and a concrete, conceptual solution. The process involves documenting the “as-is” state, identifying stakeholder needs and pain points, and then defining a “to-be” vision where GenAI can provide a transformative impact.

Key Activities

1 Analyze the “As-Is” State:

Document the current business processes, workflows, and performance metrics associated with the problem area. This includes mapping the sequence of activities, the systems used, and the roles involved.
Conduct structured interviews and workshops with stakeholders to understand their day-to-day operations, challenges, and unmet needs. The goal is to identify specific pain points and inefficiencies that GenAI could potentially alleviate.
Analyze existing data and reports to establish a baseline for current performance. This baseline will be critical for measuring the impact of the GenAI solution later.
- For example, in the wildland fire context, this would involve documenting how situational reports are currently generated or how resource allocation decisions are made during an incident.

2 Identify and Prioritize GenAI, AI, Analytics and Data Workflow Use Cases:

Brainstorm a list of potential GenAI use cases based on the “as-is” analysis. These use cases should be directly linked to the identified pain points and desired business outcomes.
Categorize the use cases into logical MISSION groups. The GenAI Wildland Fire reports categorize use cases by disaster management phase (e.g., Mitigation, Preparedness, Response, Recovery) or by stakeholder group (e.g., Helping People, Insurance Business Management).
Categorize the use cases of the capabilities and consider using a framework to guide stakeholders of what areas to consider the use cases capabilities touch:

Data & Foundation

Services & Reasoning

Applications & Interaction

[Pipelines] ingest and agent models for big, fast, geo data to feed advanced analytics and AI models.

[Foundation] models integration with Pre-trained GeoAI Models, LLM/SLM, and Domain-Specific Models (Prithvi, SAM, ClimaX)

[Review & Extraction] LLM, SLM, classification, Geospatial tagging, and extraction, real-time workflows, and from large unstructured lakes for documents, various imagery, LiDAR, feeds

[Reasoning] lab development of Task-Specific, Fine-Tuned, and integrated Generalized models for Skills, Agents, and Utilities

[Adapt and Analyze] with ML analytics, LORA, regression, transform, forecast and Foundation Model training

[Components] and Patterns for GeoAI Models, AI Skills (e.g., Code Generators), AI Agents (e.g., Data Explorer), and Utility Services (e.g., Text/Vision) including RAG and ReAct

[Integrate] (The How) analytics, skills, agents, lakes in advanced visualizations in common tools (GIS, BI, ERP, etc.)

[Visualize] (The What) with Tracking, Classification, Detection in maps, apps, BI, Digital Twins, 3D, AR, XR

[Interact] (Actions for the Who) Engage via AI Assistants for app creation, coding, and data discovery. Use Voice, Chat, Q&A Prompts, and other agent-driven exploration tools.

Evaluate and prioritize these use cases using a formal matrix. This matrix should score each use case against a set of defined criteria. As demonstrated in the OGC CDRP reports, valuable criteria include:
- Stakeholder Need / Mission Impact: How critical is this for the end-user and the organization’s mission?
- GenAI Value: How significant is the potential improvement that GenAI can provide compared to existing solutions?
- Categorization: Have capability to breakdown by the mission or AI Capability category tagged to assist in trends or to uncover findings.
- Technical Feasibility: How practical is it to develop and implement this use case with current technology and data?
- Data Readiness: Is the necessary data available and of sufficient quality?
The output should be a ranked list of use cases, allowing the team to focus on the one(s) with the highest potential for success and impact, such as “Community Risk & Resilience Assessment” or “Loss Analysis for Portfolio Management”.

3 Develop the “To-Be” Vision:

For the highest-priority use case(s), develop a clear and detailed “to-be” vision. This is a narrative and visual description of the desired future state with the GenAI solution implemented.
Create a conceptual workflow diagram that illustrates how users will interact with the new GenAI-powered process.
Consider Components needed:

Consider the full data, information, knowledge integration with AI Wisdom capabilities noted in Phase 4.

Data Aggregation	Information Platforms	Data Platforms	Knowledge Platforms
Serverless Data Pipelines Traditional ETL Batch Data Data Quality RPA Services Metadata Harvesters Web Service Processing Big Data ELT Pipelines Real-Time Data Hubs Large Feed & Raster Processing	Web, Voice, & Mobile App Dev MIS Stack Integration (e.g. MIS, ERP, CRM, etc.)	Interfaces / APIs Geospatial Platforms Business Intelligence Rules-Based & Gamified Search Signals Metadata Catalogs Enterprise Datasets	Data Science Cloud Workbenches Advanced Data Analytics Platforms Semantic Platforms Mission Data Lakes Orchestration Data Warehouse

Develop a high-level conceptual architecture diagram. This diagram should show the key components of the solution—such as Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and potential AI Agents—and how they connect with data sources and end-users. It should also illustrate the flow of data from input to a potential GenAI-generated solution. Use the reference below to help in laying out and organizing the groups of components.

Inputs for this Phase

Project Charter and Stakeholder Map (from Phase 1).
Access to subject matter experts and end-users for interviews and workshops.
Documentation of existing processes and systems.

Outputs of this Phase

“As-Is” Process Analysis Report: A document detailing the current workflows, pain points, and performance baselines.
Prioritized Use Case Matrix: A ranked list of potential GenAI use cases with their evaluation scores, justifications, and a recommendation of which use case(s) to pursue.

“To-Be” Conceptual Design Document: A document containing the narrative vision, conceptual workflow diagrams, and high-level architecture for the selected use case.