Execution Phases

Phase 1: Initialization (0:00 - 0:15)

The system establishes connectivity and plans the research strategy.

[SYSTEM]               → Connecting to use case tool...
[RESEARCHER: PLANNING] → Planning the research strategy and subtasks...
[SYSTEM]               → Establishing secure connection...

Phase 2: Initial Research (0:15 - 2:00)

The AGENT: RESEARCH begins broad research based on the submitted business context. The agent crawls the company website, analyzes publicly available information, and gathers market intelligence. The Connection Log shows elapsed time counters (10s to 120s) as research progresses.

[AGENT: RESEARCH] → Conducting initial research... (10s elapsed)
[STATUS]          → Research in progress...
[AGENT: RESEARCH] → Conducting initial research... (120s elapsed)

Phase 3: Subtopic Research (2:00 - 4:00)

The research agent decomposes the broad analysis into specific subtopics and investigates each in parallel. The Connection Log displays a raw JSON-formatted agent message with the research agent's identifier, followed by messages for each subtopic.

Typical subtopics observed for a cloud services company:

• Company's Market Positioning in the cloud/AI ecosystem
• Priority AI Automation Use Cases with quantified ROI potential
• Platform Enhancement Opportunities with competitive differentiation
• Implementation Framework and Guidelines for enterprise adoption

Two writer agents work in parallel:

AGENT: STRUCTURED_WRITER generates the AI Discovery Report using a structured template. Elapsed time: 12s to 96s. Produces ~24,087 characters of structured content including decision matrices, architecture diagrams, scoring tables, and roadmap phases.

AGENT: WRITER composes the Deep Research Report as a narrative document with sections, data tables, and citations. Elapsed time: ~50s. Produces ~55,867 characters of research narrative covering competitive positioning, industry opportunities, platform strategy, and strategic roadmap.

Phase 4: Report Writing (4:00 - 6:30)

Two specialized writer agents work in parallel to generate comprehensive reports from the research data.

Steps:

1. Research data is processed and structured
2. Two writer agents activate simultaneously:
   • AGENT: STRUCTURED_WRITER -- Generates AI Discovery Report
   • AGENT: WRITER -- Composes Deep Research Report
3. Each agent produces distinct content types with specific formatting
4. Elapsed time counters track writing progress (12s to 96s)
5. Character counts are tracked internally

AGENT: STRUCTURED_WRITER Output

• Format: Structured template with sections, tables, matrices
• Content: Decision matrices, architecture diagrams, scoring tables, roadmap phases
• Length: ~24,087 characters
• Elapsed Time: 12s to 96s

[AGENT: STRUCTURED_WRITER] → Generating AI Discovery Report... (12s)
[AGENT: STRUCTURED_WRITER] → Writing decision matrices and scoring tables... (48s)
[AGENT: STRUCTURED_WRITER] → Finalizing structured report sections... (96s)

AGENT: WRITER Output

• Format: Narrative document with sections, citations, data tables
• Content: Competitive positioning, industry opportunities, platform strategy, strategic roadmap
• Length: ~55,867 characters
• Elapsed Time: ~50s

[AGENT: WRITER] → Composing Deep Research Report... (15s)
[AGENT: WRITER] → Writing research narrative and analysis... (35s)
[AGENT: WRITER] → Finalizing report sections and citations... (50s)

(6:30 - 7:00)

The dual publisher agent publishes both reports and confirms completion with exact character counts.

Steps:

1. AGENT: DUAL_PUBLISHER receives completed reports from both writers
2. Reports are validated for completeness and format
3. Exact character counts are calculated and logged
4. Reports are published to the artifact system
5. System transitions to use case generation mode

[AGENT: DUAL_PUBLISHER] → Publishing AI Discovery Report...
[AGENT: DUAL_PUBLISHER] → Publishing Deep Research Report...
[AGENT: DUAL_PUBLISHER] → Reports published (standard: 55,867 chars, structured: 24,087 chars)
[STATUS]                → Multi-agent research completed — preparing for use case generation
[STEP: REVIEW]          → Initiating use case generation from research data...

(7:00 - 8:00)

The system streams real-time use case generation based on the completed research.

Steps:

1. Research data is processed for use case extraction
2. AI identifies priority automation opportunities
3. Each use case is generated with detailed metadata
4. ROI calculations are performed for each use case
5. Alignment scores are calculated based on business goals
6. Use cases are prioritized by impact and feasibility

Real-Time Display

• Counter Badge: Increments live ("3 use cases generated" to "6 use cases generated" to "12 use cases generated")
• JSON Preview: Displays each use case as it's created
• "View Analysis..." Button: Appears when generation completes

Each Use Case Includes

[STATUS] → Generating use cases... (3 generated)
[STATUS] → Generating use cases... (6 generated)
[STATUS] → Generating use cases... (9 generated)
[STATUS] → Generating use cases... (12 generated)
[STATUS] → Use case generation complete

Typical output: 10-12 use cases with full metadata.

Analysis modal closes and the Results Dashboard loads with the completed project.

Steps:

1. Analysis modal transitions to success state
2. Green checkmark icon appears
3. Status updates: "Saving Your Project..." to "Project Saved"
4. Auto-generated Project ID is created and displayed
5. Project ID becomes editable in input field
6. Four report tabs appear below success indicator
7. Use case grid populates with all generated use cases

Success State Components

Report Tabs

Use Case Grid

• Layout: Two-column card grid
• Card Content: Title, truncated description, impact badge
• Impact Badges:
  • High Impact (red badge)
  • Medium Impact (yellow badge)
  • Low Impact (blue badge, if applicable)
• Interaction: Click any card to view full use case details

Each entry follows: [AGENT_TYPE: SUBTYPE] → Message... (elapsed time)

Color coding:

Error Handling & Recovery

If the pipeline encounters an error during execution:

• The Connection Log displays the error message
• The timer pauses
• The Cancel button remains available to abort
• If using "Run in Background," the project is saved with partial results where possible

To retry after an error, use the "Start again" tab on the Results Dashboard to re-submit the form.