Domain Extensibility: AI Generating FinOps Expertise from Research
Watch AI study your UC1 knowledge, create domain roles/workflows, then produce $323K ROI analysis
Prerequisites
- Completed UC1 (knowledge foundation about FinOps Framework)
- Claude Code or GitHub Copilot with maenifold MCP
- Access to FinOps Hubs (optional, for live data)
Setup
Ensure UC1 research complete: memory://research/finops-framework/ exists
Built-in workflows available: role-creation-workflow, higher-order-thinking
FinOps Toolkit help files and DB schema (if working with live data)
maenifold configured as MCP server
Walkthrough Examples
Example 1: AI Creates Domain Roles from UC1 Knowledge
Scenario: AI studies FinOps Framework knowledge from UC1 and creates 3 specialist roles.
Claude Code Prompt:
"Use role-creation-workflow. Study my memory://research/finops-framework/ knowledge and the FinOps Foundation website to create three roles: finops-practitioner, cfo, and ftk-agent."
What AI Does (role-creation-workflow):
- Assumes prompt-engineer role (10/10 constitutional AI expert)
- Defines role specifications from FinOps Framework
- Researches domain knowledge (reads UC1 memory files + FinOps.org)
- Analyzes constitutional requirements (principles, anti-patterns)
- Designs cognitive architecture (workflows, evaluation criteria)
- Creates role structure (JSON with motto, principles, approach)
- Validates against prompt engineering checklists
- Saves to assets/roles/:
- finops-practitioner.json (FinOps Framework expert)
- cfo.json (Executive financial stewardship)
- ftk-agent.json (KQL query executor for FinOps Hubs)
Output:
- 3 production-ready role definitions
- Each role studies FinOps Framework knowledge from UC1
- AI generated domain expertise from external docs
Why This Matters: AI creates its own domain knowledge by studying, not pre-programming.
Example 2: AI Designs Domain Workflows Using New Roles
Scenario: AI adopts finops-practitioner role to design analysis workflows.
Claude Code Prompt:
"Adopt finops-practitioner role. Use higher-order-thinking workflow to design ftk-query and ftk-analysis workflows for Azure cost analysis."
What AI Does:
- Adopts finops-practitioner (FinOps Framework expertise)
- Uses higher-order-thinking workflow:
- Examines thinking processes for workflow design
- Evaluates cognitive approaches
- Synthesizes multiple perspectives
- Designs ftk-query (data collection) and ftk-analysis (strategic reporting)
- Then adopts ftk-agent role to add correct KQL queries to workflows
- Saves to assets/workflows/:
- ftk-query.json (FinOps data collection & optimization)
- ftk-analysis.json (FinOps strategic analysis & reporting)
Output:
- Domain-specific workflows designed by AI
- Query patterns from ftk-agent role
- Strategic frameworks from finops-practitioner role
Why This Matters: AI designs workflows by adopting domain expertise roles it created.
Example 3: Execute Workflows → Produce $323K ROI Report
Scenario: Run ftk-analysis workflow using all 3 roles to produce executive report.
Process:
1. ftk-agent executes KQL queries against FinOps Hubs
- Collects
[[cost-data]],[[commitments]],[[anomalies]] - Stores results in memory:// with [[WikiLinks]]
2. finops-practitioner analyzes findings
- Applies
[[FinOps-Framework]]best practices - Identifies
[[optimization-opportunities]] - Calculates
[[ROI]]and[[payback-period]]
3. cfo synthesizes executive report
- Strategic context for board presentation
- Risk assessment and mitigation
- Multi-scenario financial projections
Output (SONNET-A Report):
Strategic Roadmap: 18-month transformation to B+ FinOps maturity
Why This Matters: $323K savings identified → ROI on setup time achieved immediately.
The Meta-Capability
This isn't about pre-built FinOps roles. It's about:
- UC1: AI researches FinOps Framework → builds knowledge graph
- UC2: AI studies that knowledge → creates domain roles
- UC2: AI adopts roles → designs workflows
- UC2: AI executes workflows → produces $323K ROI analysis
The Innovation: AI generating domain expertise on demand by studying external docs.
Demo Artifacts - Available for Inspection
Real FinOps Analysis Outputs (SONNET-A folder)
Available on website for inspection:
- finops-strategic-report.md (367 lines, executive analysis)
- executive-summary.md (business impact)
- implementation-roadmap.md (18-month transformation plan)
- recommendations.md (actionable optimizations)
- roi-analysis.json (financial projections)
- 20+ JSON data files (cost breakdowns, forecasts, anomalies)
Domain Roles
- finops-practitioner.json (FinOps Framework principles)
- cfo.json (Executive financial stewardship)
- ftk-agent.json (KQL query executor)
Domain Workflows
- ftk-query.json (Data collection methodology)
- ftk-analysis.json (Strategic reporting workflow)
Code Sample
# Create roles from research knowledge
maenifold --tool Workflow --payload '{
"workflowId": "role-creation-workflow",
"response": "Create finops-practitioner role from memory://research/finops-framework/"
}'
# Design workflows using new roles
maenifold --tool Workflow --payload '{
"workflowId": "higher-order-thinking",
"response": "Design ftk-analysis workflow for Azure cost optimization"
}'
# Execute analysis (requires FinOps Hub access)
maenifold --tool Workflow --payload '{
"workflowId": "ftk-analysis",
"response": "Analyze Azure costs for fiscal year 2025"
}'ROI Calculation
Setup Investment:
- UC1 Research: 60-90 minutes (one-time)
- Role Creation: 60-90 minutes (3 roles, one-time)
- Workflow Design: 30-45 minutes (one-time)
Total Setup: ~3 hours one-time investment
Output Value:
- $323K annual savings identified
- 638% ROI
- 1.9-month payback period
- Ongoing analysis capability
Why Should I Care? Configure once, get ongoing multi-perspective financial analysis forever. Suddenly you have analyst + practitioner + CFO analyzing your spend.
Common Pitfalls
Next Steps
- →UC3: Single-Agent Dev Workflows
Benefit without multi-agent complexity
- →
Try applying this pattern to different domains
Security, compliance, architecture—any domain you can research
- →Explore Tools & Workflows
Browse workflow tools and memory management capabilities