Bonus 12 — Innovation & Organisational Improvement

60-Second Summary

Bonus module 12 of the Manager-of-Managers program. Theme: Improve the system, not just the output.
Improvement backlog + first experiment — the real artefact you produce.
Same shape as core 12: 90-min pre-read, 4-hr monthly intensive, falsifiable artefact.
Reviewed by CHRO, VP/Director, sitting CEO, and OB faculty lenses.

Most directors out-execute the system they inherit. The leaders who scale redesign it. The discipline isn't transformation theatre — it's a continuous improvement cadence: hypothesise, run a small experiment, measure, scale or kill. This module installs that cadence and the specific 2026 lens of AI as a system-redesign tool, not a productivity bolt-on.

What the evidence says

Toyota Production System (Womack, Jones — Lean Thinking): continuous improvement compounds; one-shot transformations rarely outperform sustained kaizen.
Edmondson — The Right Kind of Wrong: organisations that run more small experiments out-learn those that run fewer big ones, regardless of strategy quality.
MIT Sloan / BCG 2024 AI adoption research: the value gap is not the tool — it's the workflow redesign. Functions that redesign earn 2–4× the productivity gain.

Pre-read (90 minutes)

Read: Lean kaizen — small daily improvements at the team layer (20 min).
Read: Edmondson on intelligent failure and the right kind of wrong (20 min).
Read: workflow redesign with AI — the 4 task classes (judgement / context / pattern / volume) and which AI changes (20 min).
Reflect (30 min): list 5 things in your function that haven't been redesigned in 2 years. Why not?

Monthly intensive (4 hours)

Cohort flow with a senior practitioner coach

1
Improvement backlog (45 min)
Each leader writes 10 candidate system improvements (process, tool, workflow, automation). Coach scores them on impact × effort.
2
Experiment design (60 min)
Pick one. Design as an experiment: hypothesis, metric, sample, duration, scale-or-kill criteria. Coach pressure-tests — most experiments are too big and too long.
3
AI redesign drill (60 min)
Take one team workflow. Classify each step (judgement / context / pattern / volume). Identify which steps AI changes and how. Design the redesigned workflow.
4
Improvement cadence (30 min)
Design the cadence: monthly improvement review, quarterly experiment readout. Coach demonstrates how to make this lightweight enough to survive a busy quarter.
5
Wrap (45 min)
Public commitment: one experiment running by end of month; one AI-redesigned workflow shipped this quarter.

The artefact you produce

Improvement backlog + first experiment

A live backlog of system improvements (impact × effort scored) and one running experiment with hypothesis, metric, and scale-or-kill date. Reviewed monthly.

Tools at this layer

Layer	Examples (2026)	Use
Improvement methodology	Lean kaizen, PDCA cycles, A3 problem-solving, retro-derived improvements	Frameworks for small, repeated change
Experimentation	Optimizely / GrowthBook for product; experiment templates for ops; pre-registered hypotheses	Make experimentation cheap and visible
AI tooling	Claude / ChatGPT / Gemini / Copilot for general; Notion AI, Linear AI, Lattice AI for workflow	Tool selection follows workflow redesign, not the other way around
Automation	Zapier, Make, n8n, Workato, internal scripts	Eliminate volume work that doesn't need humans

Copy-paste AI prompt

Here's a workflow in my function [paste the steps with time/effort estimates]. Help me: (1) classify each step as judgement / context / pattern / volume, (2) identify which steps AI changes most, (3) design a redesigned workflow with AI in the loop and human checkpoints, (4) suggest a 4-week experiment to test the new workflow with success criteria.

Between-session homework

Improvement backlog with 10+ scored candidates.
One experiment designed and running with named hypothesis and metric.
One workflow redesigned with AI in the loop and shipped to one team.
Monthly improvement review on calendar.

Success signal

By end of this module, your function has a live improvement backlog, you have at least one experiment running with scale-or-kill criteria, and you've redesigned at least one workflow around AI rather than bolting AI onto an unchanged workflow.

Reviewer notes

CHRO (20+ yrs)

The leaders who improve the system create capacity that no headcount request can match. Improvement is the highest-leverage capacity investment most directors never make.

VP / Director (15+ yrs, 3+ scaled orgs)

I spent my first year as VP improving processes I'd complained about as a director. The compound effect on my team's output was larger than any hire I made.

Sitting CEO

I look for the senior leader who tells me 'we shipped 20% more this quarter with the same team because we redesigned X'. That leader gets the next role.

OB / HR Professor (25+ yrs)

The Toyota and Deming evidence base on continuous improvement is 70 years deep and still under-applied in knowledge work. The newer AI-redesign literature simply extends it: workflow change is the moat, tool adoption is not.

Written by Pawan Joshi.Sources cited inline.

First published 30 Jun 2026See site changelog →