Optimize Workflows for Peak Productivity

Workflow Optimizer is an autonomous agent that analyzes existing human-agent collaboration patterns, identifies bottlenecks and inefficiencies, and designs and recommends optimized workflows to maximize productivity and quality outcomes. It runs a four-step process: workflow discovery, examining documentation, chat logs, and project files to map human touchpoints, agent interactions, handoffs, dependencies, and current time and resource investment; bottleneck analysis, identifying delays, redundancies, context-switching overhead, communication friction, task-complexity and agent-capability mismatches, and knowledge-transfer gaps; optimization design, parallel-processing opportunities, standardized templates, handoff protocols with validation checkpoints, and escalation paths for edge cases; and implementation planning, prioritizing changes by impact versus effort, a migration guide, feedback loops, and success metrics. It uses Read, Glob, Grep, and WebSearch tools and runs on the Sonnet model.

Use it when you have an existing human-agent workflow - documented processes, chat logs, or project files showing how humans and agents currently hand work back and forth - and want a structured analysis of where it is inefficient plus a concrete, phased plan to fix it, including a decision matrix for which tasks should be fully automated, human-agent collaborative, human-led, or human-driven based on how routine, complex, critical, or novel they are. It is an analysis and design agent, not an implementation agent - it produces the optimized workflow design and migration guide, not the automation itself.

Input is existing workflow documentation, chat logs, and project files describing current human-agent collaboration. Output is a Current State Analysis (workflow map, ranked pain points with impact assessment, resource-usage breakdown by task type), an Optimized Workflow Design (new process flow, parallel opportunities, a template library of standardized formats and prompts, handoff protocols), and an Implementation Guide (a phased 1-3 rollout, success metrics, risk mitigation, feedback mechanisms), plus a reusable Agent Task Specification template covering task, context, success criteria, escalation triggers, and output format, and a decision matrix mapping task types - routine/structured, complex/creative, critical/high-risk, learning/novel - to the right level of automation versus human involvement.

Runs with Read, Glob, and Grep for examining local documentation and project files, plus WebSearch for external research, on the Sonnet model. No external project-management or automation-platform integration is documented; its output is a written analysis and design, applied by the team.

Teams running human-agent collaborative workflows who want a structured audit of where handoffs, context-switching, or capability mismatches waste time, plus a phased, measurable plan, including reusable task-specification templates and a decision matrix, for improving the process rather than just individual task performance.