M1 · THE ECONOMIC CASE · P01

The ROI Playbook

Why half see zero. Why half see 171%.

12 SLIDES · ~10 MIN · PLAYBOOK OVERVIEW
Key Takeaways
  • Half saw zero. Half saw 171%. They measured different things.
  • 56% — of CEOs see zero ROI on generative AI.
  • 171% — ROI reported by agentic deployers.
  • 88% — of companies see no return at all.
  • One ROI metric per AI initiative — before any spend.
PROGRAM DEPTH MAP

7 modules. 21 playbooks. 1 system.

M1 The Economic Case
The ROI Playbook
The Procurement Playbook
The Model-Swap Playbook
M2 First Functions
The Sales Growth Playbook
The Procurement Playbook
The Agent Migration Playbook
Function Before Industry
M3 Company Brain
The Company Brain Playbook
The Discoverability & Governance Playbook
The Model-Swap Playbook
M4 Prototype to Production
The Prototyping Loop Playbook
The Hidden Cleanup Playbook
The Anthropic Agent Environment Playbook
M5 Organization Rewrite
The Workforce Tiering Playbook
The Org Change Playbook
The Campfire Protocol Playbook
M6 The 90-Day Playbook
The 90-Day Capstone Playbook
The Stack Inheritance Playbook
M7 How I Run My Day
The Chief-of-Staff Playbook
The Build Playbook
The Security Stack Playbook

This is not an AI overview. A sequenced operating system for Monday-morning execution — each playbook is a single thesis, executed Monday morning.

WORKSHOP TRACK

11 workshops. Run Monday.

WS·01
The Secrets Workshop
WS·02
Enterprise AI Memory Architecture
WS·03
Clean Zone Architecture
WS·04
GitHub for CEOs
WS·05
The AI Grifter Test
WS·06
CloseBot — AI Lead Qualification
WS·07
From Conversation to Capability
WS·08
The Resolution Layer — Why Your AI Gets Confused
WS·09
Windows Just Became an Agentic OS
WS·10
Trustworthy Memory — Resolution Deep Dive
WS·11
The AI Portability Playbook

Workshops · CEO deep-dives. Each one is a 30-minute working session a CEO can run with their team on Monday morning.

MODULE 1 · PLAYBOOK 01

The ROI Playbook

Why half see zero. Why half see 171%.

BuildClub Academy
01 · WHY THIS MATTERS

Same companies, same tools, opposite outcomes.

56%
of CEOs see zero ROI on generative AI.
PwC 2026
171%
ROI reported by agentic deployers.
Microsoft Wave 1 / 2026
88%
of companies see no return at all.
Wharton AI ROI 2026
02 · THE THESIS

Half saw zero.
Half saw 171%.
They measured different things.

The Thesis
03 · THE BIMODAL SPLIT

The split is measurement, not capability.

Zero-ROI camp 171% camp PwC: zero ROI 56% of enterprises 56% MIT NANDA: pilots stall 95% of pilots 95% Wharton: positive ROI 74% outcome-based 74% Agentic deployers 171% ROI (indexed 100) 171% Same companies. Same tools. Opposite outcomes.
Takeaways
  • 1 PwC: 56% of enterprises report zero ROI on generative AI.
  • 2 Wharton: 74% report positive ROI when measurement is outcome-based.
  • 3 MIT NANDA: 95% of pilots stall — almost all lack a named metric on day one.
04 · THE PLAYBOOK

The 5-step ROI Playbook.

1 Name the metric the kill-or-scale variable 2 Pick the function one function, not a portfolio 3 Pick a platform buy-the-platform test passes 4 Run shadow mode human + agent in parallel 5 Cut what can't be measured
05 · WHAT TO MEASURE

Outcome metrics vs activity metrics.

Activity (default)
Seats deployed · Prompts written · Tokens consumed · Training hours.
Outcome (correct)
$ per resolution · $ per closed deal · $ per decision · Hours saved per workflow.
What vendors push
Seat licenses, usage tiers, training credits. Optimized for their P&L, not yours.
What boards ask
Where does this show up on the income statement? Activity metrics can't answer that question.

Vendors that won't sign outcome contracts don't believe their own product.

06 · WHERE THE VALUE COMES FROM

BCG 10-20-70: 10% tech, 20% process, 70% people.

BCG AI at scale — where the value comes from 10% Tech 20% Process 70% People Source: BCG AI at Scale 2026 — sized across 1,400+ enterprise deployments.
Takeaways
  • 1 Companies budgeting 80% on tech and 5% on people are budgeting for the zero column.
  • 2 The 70% is fluency, champion networks, change management — not training videos.
  • 3 Audit the last 10 AI purchase orders: how much went to people enablement?
07 · WHY PILOTS STALL

Four named failure modes.

1
No metric named at the start
You can't be wrong if you can't be measured. You also can't be right.
2
Activity reported as outcome
Training hours and prompts written are not throughput. The board can tell.
3
Sponsor without budget authority
The AI hire problem: someone champions, no one funds. Pilot starves.
4
No shadow-mode baseline
You don't know what 'better' looks like without running the old way alongside.
08 · THE 171% CAMP DOES THIS

20 agents, 1.2 humans, $2.4M closed.

1
Outcome metric: revenue per agent cell
Not seats. Not tokens. Closed pipeline divided by agent count.
2
Software bill: under $5K/month
The platform was the cheapest input. The metric did the work.
3
Result: $2.4M closed pipeline
Eight reps compressed to 1.2 FTE-equivalent. Same revenue. Different cost structure.
CLOSING

Five things to do Monday.

Monday Morning
  1. 1
    One ROI metric per AI initiative — before any spend.
  2. 2
    No purchase order without a metric attached.
  3. 3
    Inventory the last 10 AI vendor pitches: agentic vs generative.
  4. 4
    Define one outcome KPI per AI initiative — refuse to fund the next without the prior having one.
  5. 5
    Reallocate AI budget toward the BCG 10-20-70 split.
P01 · Questions CEOs Ask

Frequently Asked Questions

What is the core idea of The ROI Playbook?
Half saw zero. Half saw 171%. They measured different things.
What does the data say a CEO should pay attention to?
56% of CEOs see zero ROI on generative AI. 171% ROI reported by agentic deployers.
What should a CEO do Monday morning after reading The ROI Playbook?
Start here: One ROI metric per AI initiative — before any spend; No purchase order without a metric attached; Inventory the last 10 AI vendor pitches: agentic vs generative.
Where do the claims in The ROI Playbook come from?
The playbook cites BuildClub Academy program architecture; BuildClub Academy workshop catalog; PwC 2026 AI Predictions; Microsoft Release Wave 1 / 2026.
P01 · Sources

All Sources in This Playbook

00 / 12