Hi, I'm Andy Pray.
I am interested in the full lifecycle of governed enterprise agentic AI — choosing and funding the right bets, shipping reliable agent products, and building the organization's ability to adopt and scale them.
I have set technology strategy for a 1,500-person compliance organization and architected multi-agent systems on LangGraph supervisors and explicit state machines, with human-in-the-loop approvals, deterministic guardrails and SOC 2 / ISO 27001 alignment.
I help teams bridge strategy and execution—translating value cases into governed, evaluated, and observable agent products. Select a topic below to explore the skills and technologies that enable this work.
Where should we place AI-agent bets, and how will they create measurable value?
AI Opportunity Strategy
Skills
- Problem framing
- Customer and employee workflow analysis
- Use-case discovery
- Value-proposition design
- Competitive and technology assessment
- Telling automation, copilot, agent and multi-agent apart
Technologies and Methods
- Opportunity maps
- Jobs-to-be-done
- Workflow decomposition
- Service blueprints
- Agent-suitability assessments
- Capability maps
How do we build, deploy, evaluate and operate trustworthy agents?
Agent-System Architecture
Skills
- Agent decomposition
- Single-agent versus multi-agent design
- Planner and executor patterns
- Workflow and state-machine design
- Tool-use design
- Bounded autonomy
- Failure-mode analysis
Technologies and Methods
- LangGraph
- Explicit state machines
- Supervisor patterns
- ReAct
- Tool calling
- Google ADK
- MCP
- Python
How do we make AI capabilities change the organization at scale?
AI Transformation Operating Model
Skills
- Enterprise capability design
- Role definition
- Decision-rights design
- Product operating models
- Central-versus-federated team design
- Platform governance
Technologies and Methods
- Target operating models
- RACI and DACI
- AI center-of-excellence models
- Federated product-team patterns
- Reusable reference architectures
