Skip to content

Latest commit

 

History

History
38 lines (27 loc) · 2.29 KB

File metadata and controls

38 lines (27 loc) · 2.29 KB
title McKinsey — Workflow Redesign First
source-id mckinsey-state-ai-2025
wiki-page (02_References/enterprise-ai/wiki/skills/process/redesign-workflows-around-ai.md)
last-synced 2026-06-07

McKinsey — Workflow Redesign First

What this reference contains

McKinsey State of AI 2025 findings on the workflow redesign imperative: AI high performers are 3× more likely to fundamentally redesign workflows before deploying AI. The failure mode — AI bolted onto a legacy workflow — is the single most common cause of marginal AI returns.

Key claims (with citations)

  • 3× workflow redesign. AI high performers are 3× more likely to fundamentally redesign workflows before deploying AI vs. the average firm. Source: McKinsey State of AI 2025.
  • AI bolted onto legacy workflow = marginal benefit at best. The workflow is the intervention; AI is the accelerant. Applying AI to a broken process scales the broken process. Source: McKinsey State of AI 2025 + Stanford AI Index 2026.
  • The redesign question to ask: "If we had no AI at all and were designing this workflow from scratch, how would we design it?" The answer to that question is the target state. Then: "Where does AI change this design?"
  • Process redesign before AI deployment reduces the total-cost-of-deployment significantly (fewer failed integrations, less rework, lower change management cost). Source: McKinsey.
  • High performers change WHAT the process does, not just how fast it runs. Speed-up of an existing process is efficiency. Changing the process design is value-creation.

The workflow redesign checklist

Before AI deployment, answer:

  1. What is this workflow trying to accomplish?
  2. Which steps exist because of constraints that AI removes?
  3. Which steps add value and should be kept?
  4. Which steps can now be eliminated entirely?
  5. What does the redesigned workflow look like? (Draw it before deploying)

When to consult

  • For skill process-pilot-design: workflow redesign is step 0 of pilot design — before any technology decisions.
  • For skill tech-stack-diagnostic: if a stack diagnostic finds AI is deployed without workflow redesign, this is a process-layer gap, not a technology gap.

Source

McKinsey State of AI 2025. 3× workflow redesign finding; AI bolted onto legacy workflow failure mode.