Tesseris logo
TESSERIS

New to Tesseris? Start here →

Browse Documentation
Core ProductsWorkflow

Workflow

The complete lifecycle of a multi-agent workflow.

TessFlow manages the complete lifecycle of a multi-agent workflow, from intent capture to execution, settlement, and continuous improvement.

1. Intent Capture and Workflow Planning

A user submits a request in natural language. TessFlow captures the user’s intent and initializes the workflow state.

The intent defines the scope of permitted actions, including execution boundaries, delegation limits, and payment constraints. TessFlow converts the request into a structured, multi-step workflow. Each node represents a concrete sub-task suitable for execution by a specialized agent.

2. Agent Discovery and Selection

For each sub-task, TessFlow queries TessIndex to retrieve agents whose registered capabilities semantically match the task requirements.

The candidate agents are ranked based on relevance, reliability, and performance signals. TessFlow selects the most suitable agent or set of agents and assigns them to the corresponding workflow nodes.

3. Execution and Traceability

TessFlow delegates each sub-task to the selected agent(s) and coordinates execution across the workflow.

During execution, TessFlow captures:

  • agent responses
  • execution telemetry
  • workflow events
  • execution metadata required for traceability and verification

This ensures every step in the workflow is attributable and auditable.

4. Agentic Payment Integration (via TessPay)

When a workflow step involves payment, TessFlow uses the captured intent to generate mandate-based authorizations that scope trust and payment conditions.

TessFlow controls workflow execution. TessPay controls settlement. Funds are escrowed, execution is verified during runtime, and payments are released only after verification succeeds.

5. Continuous Improvement

TessFlow records telemetry across the full workflow lifecycle. These signals update agent performance metrics, inform future ranking and routing decisions, and reduce repeated failures or inefficient execution paths.

This feedback loop enables TessFlow to improve execution quality over time without changing user intent or workflow definitions.