Advisory Engagement Tool Dispatch and Routing Matrix
Structure multi-tool routing rules and execution dependencies for management consulting AI agents.
Use this template when configuring multi-tool workflows and routing logic for client-facing advisory agents. It produces a clear dispatch matrix detailing triggers, execution constraints, and human-in-the-loop gates.
Role: Senior Enterprise AI Integration Consultant specializing in multi-agent tool orchestration.
Context
- Client Name: {{consultancy_client}}
- Engagement Scope: {{engagement_type}}
- Connected Enterprise Systems: {{erp_crm_sources}}
- Orchestration Framework: {{agent_framework}}
- Human Approval Trigger: {{human_in_the_loop_threshold}}
- Retry & Fallback Policy: {{error_handling_policy}}
Task
Create an Advisory Tool Dispatch and Routing Matrix for {{consultancy_client}} that defines deterministic routing criteria, tool dependencies, and human approval boundaries for {{engagement_type}} workflows.
Method
- Analyze {{erp_crm_sources}} to identify read and write tool functions.
- Map intent triggers from user queries to specific function calling schemas in {{agent_framework}}.
- Establish sequential and parallel execution dependencies between tool calls.
- Apply {{human_in_the_loop_threshold}} criteria to flag sensitive tool invocations.
- Define validation criteria for intermediate tool outputs before downstream dispatch.
- Integrate {{error_handling_policy}} into tool retry and rollback states.
- Format all routing logic into an actionable dispatch decision matrix.
Constraints
- MUST flag any destructive or write operations for human approval under {{human_in_the_loop_threshold}}.
- MUST NOT leave output-parsing dependencies ambiguous between chained tools.
- Include explicit error response handling for all integrated systems.
- Keep matrix definitions concise and actionable.
Output format
1. Routing Architecture Summary
A single paragraph outlining the dispatch flow within {{agent_framework}}.
2. Tool Dispatch Matrix
A markdown table with 7 columns: | Tool Identifier | Target System | Trigger Intent | Input Arguments | Output Format | HITL Gate Required (Yes/No) | Fallback Action |
3. Orchestration Directives
Exactly 3 bulleted instructions for managing state across chained tool calls.
Self-review
- Ensure all systems in {{erp_crm_sources}} have corresponding matrix entries.
- Check that HITL requirements match {{human_in_the_loop_threshold}}.
- Verify error fallback steps align with {{error_handling_policy}}.
Explicit role, a named task, and discrete steps the model can follow.
Background, inputs and variables the model needs before it starts.
Hard boundaries — what the model must and must not do.
A named, field-level shape for the response.
Ordered work items that force analysis before an answer.
Length and structure that travel across frontier models.
Signal density — instruction weight without padding.
Documented variables so the scaffold adapts to new inputs.
Quality bar, assumptions and behaviour when inputs are thin.
How much real usage the template has behind it.