Corcava logoLe seul outil métier dont vous avez besoinCorcava
Menu
Solutions
Secteurs
DesignersRH & recrutementCabinets d'avocatsImmobilierConseilMarketing digitalComptabilitéE‑commerceDéveloppement logiciel
Produit
FonctionnalitésIntégrations
BlogTélécharger
Ressources
TemplatesDocsMCPFAQAPIComparaisonsFeuille de routeQuoi de neuf$ Programme d'affiliation
Tarifs
Essai gratuit de 14 jours

Aucune carte bancaire requise

Goulots d'étranglement avec MCP : repérer où le travail bloque

Identify workflow bottlenecks by analyzing boards and task throughput using MCP. This guide shows you how to analyze board columns, detect where tasks get stuck, and generate actionable recommendations—all without changing any data.

Ce que ce workflow permet

Board bottleneck analysis with MCP reveals workflow issues:

Résultats clés

  • Board analysis: Examine board structure and column states
  • Bottleneck detection: Identify columns where tasks accumulate
  • Throughput analysis: Analyze task flow and velocity
  • Actionable recommendations: Get suggestions to improve flow
  • Read-only insights: All analysis without modifying data

Prérequis

Before using this workflow, ensure you have:

Workflow pas à pas

Step 1: Get Board Structure

Start by understanding the board layout:

Board Analysis Prompt

"Analyze board [Board Name or ID] for bottlenecks: - Get board details: columns, states, and structure - Show me the board layout and column flow - Identify the typical task progression path"

What the AI does:

  1. Calls get_board to get board structure
  2. Analyzes columns and states
  3. Maps typical task flow through columns
  4. Presents board structure

Step 2: Analyze Task Distribution

See where tasks are accumulating:

Distribution Analysis Prompt

"Analyze task distribution across board columns: - Count tasks in each column/state - Identify columns with unusually high task counts - Find columns where tasks stay longest - Show me the distribution breakdown"

What the AI does:

  1. Calls list_tasks filtered by board
  2. Groups tasks by column/state
  3. Counts tasks per column
  4. Identifies columns with high task counts

Step 3: Detect Bottlenecks

Identify where work gets stuck:

Bottleneck Detection Prompt

"Detect bottlenecks in the board: - Find columns with many tasks but slow movement - Identify columns where tasks accumulate (more coming in than going out) - Flag columns with tasks stuck for long periods - Analyze task age in each column Show me which columns are bottlenecks and why."

What the AI does:

  1. Analyzes task counts per column
  2. Calculates task age in each column
  3. Identifies accumulation patterns
  4. Flags bottleneck columns

Step 4: Analyze Throughput Signals

Examine task flow patterns:

Throughput Analysis Prompt

"Analyze task throughput: - Calculate average time tasks spend in each column - Identify columns with slowest throughput - Find patterns: which columns consistently slow down flow - Compare column throughput rates Show me the throughput analysis."

What the AI does:

  1. Analyzes task status change history
  2. Calculates time-in-column metrics
  3. Identifies slow columns
  4. Compares throughput across columns

Step 5: Generate Recommendations

Get actionable suggestions (read-only):

Recommendation Generation Prompt

"Based on the bottleneck analysis, propose actionable recommendations: - For each bottleneck column, suggest improvements - Recommend workflow changes (e.g., split column, add WIP limits) - Suggest process improvements (e.g., faster reviews, clearer criteria) - Prioritize recommendations by impact Important: Only provide recommendations, don't change any data."

What the AI does:

  1. Analyzes bottleneck causes
  2. Proposes workflow improvements
  3. Prioritizes recommendations
  4. Presents actionable suggestions (read-only)

Complete Workflow Prompts

Here are complete, copy-paste ready prompts:

Full Bottleneck Analysis

"Analyze board [Board Name] for bottlenecks: 1. Get board structure and column layout 2. Analyze task distribution across columns 3. Detect bottlenecks (columns with accumulation) 4. Analyze throughput (time in each column) 5. Generate actionable recommendations Important: Only analyze and recommend, don't change any data."

Use this when: You want a complete bottleneck analysis

Quick Bottleneck Check

"Find bottlenecks in board [Board Name]. Show me which columns have the most tasks and which are slowest."

Use this when: You want a quick overview

Workflow Variants

Variant 1: WIP Limit Analysis

Check if WIP limits would help:

"Analyze board for WIP (Work In Progress) limit recommendations: - Identify columns that consistently have too many tasks - Suggest appropriate WIP limits for each column - Explain how WIP limits would improve flow Focus on columns that could benefit from limits."

What happens: AI provides WIP limit recommendations

Variant 2: Review Column Analysis

Focus on review/approval bottlenecks:

"Analyze review columns (e.g., 'In Review', 'Awaiting Approval'): - Find review columns with slow throughput - Identify tasks stuck in review - Suggest process improvements for faster reviews - Recommend review time targets Focus on approval/review bottlenecks."

What happens: AI focuses on review process bottlenecks

Example: Complete Bottleneck Analysis Session

Here's a real example of how a bottleneck analysis session flows:

User Prompt

"Analyze the 'Development' board for bottlenecks. Show me where work gets stuck and recommend improvements."

AI Response (Bottleneck Analysis)

Board Bottleneck Analysis - Development Board

Task Distribution:

  • Backlog: 45 tasks
  • In Progress: 12 tasks
  • In Review: 28 tasks ⚠️
  • Done: 5 tasks

Bottlenecks Detected:

  • In Review column: 28 tasks (56% of active tasks)
    Average time in column: 5.2 days
    Issue: Tasks accumulating faster than they're being reviewed

Recommendations:

  • High Priority: Add WIP limit of 15 tasks to "In Review" column
  • Medium Priority: Split "In Review" into "Code Review" and "QA Review"
  • Process: Set review time target of 2 days maximum

Safety Patterns

✅ Read-Only Analysis

This workflow is designed to be read-only:

  • No data changes: Only reads board and task data
  • Analysis only: Provides insights and recommendations
  • Safe to run: No risk of modifying tasks or boards

Troubleshooting

No Bottlenecks Found

Symptom: AI reports no bottlenecks

Possible causes:

  • Board actually has good flow
  • Not enough task history
  • Analysis thresholds too high

Fix: Lower thresholds: "Flag columns with more than 10 tasks" or analyze specific time period

Recommendations Too Generic

Symptom: Suggestions are vague

Possible causes:

  • AI lacks context about team processes
  • Board structure is unclear

Fix: Provide context: "Our team does code review then QA. Focus recommendations on that process."

Related Tools

This workflow uses these Corcava MCP tools:

get_board

Get board structure and columns

list_boards

Find boards to analyze

list_tasks

Analyze task distribution

Related Use Cases

Risk Review

Include board bottlenecks in risk analysis

Status Reports

Include bottleneck insights in reports

Identify Workflow Bottlenecks

Connect your AI assistant to Corcava and find where work gets stuck

Get Started →Tool Reference