To Succeed with AI, Modernize Your Lean Toolkit

Recently Harvard Business Review ran an article titled "To Succeed with AI, You've Got to Nail the Basics".

The argument: instead of worrying about whether you are adopting AI fast enough, focus on getting five principles right - customer centricity, focus on process, commitment to acting on facts, a philosophy of continuous improvement, and recognition that quality happens through people.

Read that list again. Notice anything?

Those are Lean principles. Relabelled, renumbered, repackaged for a business audience - but Lean principles.

• Customer centricity = Voice of Customer
• Focus on process = Value stream thinking
• Commitment to acting on facts = Go and see - genchi genbutsu – and empiricism
• Continuous improvement = Kaizen
• Quality through people = Respect for People (one of two pillars of TPS)

The quiet conclusion from a Harvard Business Review article is that the foundation organizations need for successful AI adoption is the same foundation they need for basically everything else as well. Lean has been that foundation since Taiichi Ohno and Deming in the 1950s and 60s. Now AI is the latest thing on top of it.

This has been the thread running through every post in this series.

• Post 1: Eliminate before automating
• Post 2: Ask redesign questions before the AI question
• Post 3: Measure where AI saves time in structured improvement work - avoid the cost of hesitation / slow work
• Post 4: Define clear standards before building an agent
• Post 5: Understand the highest value applications of AI using the Priority Framework tiers
• Post 6: Avoid the investment trap of protecting bad processes with technology
• Post 7: Point AI at the bottleneck first
• Post 8: Analyze and commit to where AI-freed effort will be deployed
• Post 9: Validate every AI output - your name goes on the second draft
• Post 10: Start with the six things AI does well in your bottleneck functions
• Post 11: Pay down process debt before AI compounds it

All of it rests on one foundation:

You have to know how to improve a process without AI before AI is truly useful to you.

Organizations without that foundation will spend a lot of money and end up doing the wrong things faster.

Organizations with that foundation can use AI to expand value-adding work they previously lacked the capacity to do. That is the difference between AI as an amplifier of good process and AI as expensive noise.

Which brings me to the point that has been implicit in every post: this is why Lean training still matters.

Not because it is traditional. Not because it is a credential.

Not because it looks good on a resume.

Because it is the discipline that makes technology investment pay off.

McKinsey's 2026 analysis of AI skill gaps in operations found that only about one-third of surveyed manufacturing companies had scaled any AI solutions across their networks, and the number one reason cited for the gap was talent skill gaps. Not budget. Not tools. Skills.

The skills McKinsey is pointing at are, in my experience, Lean skills with a digital layer:

• Understanding a process well enough to know where AI fits.
• Knowing how to measure whether an intervention actually worked.
• Respecting the people doing the work enough to involve them in the redesign.

These are exactly the capabilities our top-performing clients have been building for years - long before AI came along.

If you have been building Lean capability in your organization, keep going. You are building the foundation AI needs.

If you haven't started, the fact that AI is now everywhere is the best argument for starting with Lean now - not in spite of the AI moment but because of it.

The organizations that will get the most from AI in 2026 and beyond are the ones that already know how to improve a process without it. The rest will be building chatbots on top of broken forms, writing prompts for processes that shouldn't exist, and wondering why their AI investment isn't paying off.

But what about technical skills like data engineering, prompt engineering, model evaluation, or agentic system design? Fair point. Lean is the foundation, not the entire stack.

The argument is not that Lean replaces those skills. It is that without Lean thinking, those skills get applied to the wrong problems. New technical capability only adds value when it is built on the right foundation.

Nail the basics first. Then nail them again. Then add AI - where it genuinely helps.

Key takeaways:

• HBR's five principles for AI success are just Lean principles in a different vocabulary.
• Organizations that know how to improve a process without AI get the most value from AI.
• McKinsey: the top reason AI projects fail is talent skill gaps, not budget or tools.
• Lean capability is the foundation skill AI needs - not the legacy it replaces.
• Nail the basics. Then add AI where it genuinely helps.