We Invested in AI. Nothing Changed.
Five decisions that tell you whether to fix the work, redirect the tool, pause for evidence, or stop.
You approved the pilot. The demo was strong. The team used it. Now the board asks what it returned, and you do not have an answer that holds.
We Invested in AI. Nothing Changed. A 12-page executive guide. No theory. Five decisions, one verdict.
Run the screen on one deployment
Pick one stalled or uncertain deployment. Answer each question only where you have evidence. A missing answer is a finding.
Your preliminary read:
The count is not the go decision. The screen tests work-design readiness. It does not establish a verified root cause. If the evidence points to model reliability, data, integration, or economics instead, investigate those causes. Do not force a work-design diagnosis.
Your guide author
The author behind the decisions
Dr. Ravinder Tulsiani
This guide is extracted from the decision sequence in Solve The Right Problem: five decisions, in order, that turn a stalled deployment into a funding decision. The methods behind them are CLEAR for the outcome, the Domino Map and the Performance Blueprint for the work, and IMPACT for the evidence.
Recognize the pattern
Three ways a deployment stalls. Use these as prompts for investigation, not as proof of the root cause.
A. The Capacity Trap. AI shortens one step. The queue downstream does not move. First check: handoffs, approvals, and who owns the complete flow.
B. The Governance Bypass. The sandbox works. Production approval stops the rollout. First check: authority, acceptable use, risk controls, and evidence of testing.
C. The Activity Report. Usage rises. The business result is unchanged or unknown. First check: baseline, changed behaviour, cost, and measurement design.
What each case proves, and what it does not
The bank, 2025. A bank cut 45 Direct Banking roles tied to a voice bot, then reversed the cuts within weeks after its own review found the assessment had not adequately considered the relevant business considerations. What it proves: the tool was not the main failure; the decision behind it was. What it does not prove: that every voice-bot deployment fails the same way.
The chatbot, 2024. In Moffatt v. Air Canada, a British Columbia tribunal held the airline responsible for misleading fare information provided through its chatbot. What it proves: the organization remains responsible when AI acts. What it does not prove: a universal requirement to manually approve every AI output.
The limited trial. A committee he chaired evaluated Class, a third-party add-on for engagement inside Teams. Two departments asked for it. The committee ran it through its filters, ran a risk assessment with security and technology members involved, and approved a limited trial on two computers only, for planned events. After the trial, the departments found it clunkier than expected. It added no value and created frustration. It was never rolled out broadly. It was canceled. What it proves: a bounded decision keeps the blast radius small. What it does not prove: that every limited trial ends in cancellation.
Get the guide. US$0, free.
The five decisions in full, the patterns, and the measurement contract. Five questions on the page told you where the gaps are. The guide tells you what to do about each one.
Get the guideUS$0, free.Who this is built for
This is for you if:
- ✓You are a COO, CHRO, CIO, or transformation leader who funded an AI deployment and now has to show what it returned
- ✓A live deployment is on your desk with an accountable business owner, and the next spend decision is yours to make
- ✓You want the diagnosis to carry the funding decision, not the demo
This is not for you if:
- ✕You are shopping for a tool. Nothing in this guide selects a vendor
- ✕You want a maturity score. These five decisions make an investment decision; they do not award a score
- ✕You want a guaranteed return. The guide grades readiness and supports a preliminary verdict. It does not certify a result
Pick one deployment. Decide what to do next.
Do not approve another rollout on usage data alone. Bring the owner, the workflow, and the evidence into the same discussion.
Send Me the Guide
- Full name: your name
- Work email: your work email
- What you get: the 12-page guide, “We Invested in AI. Nothing Changed.”, as a secure download link emailed to you. No attachment.
- What it costs: free
- What happens after you send this: the guide is emailed to you. The success message reads: “Check your email. The guide is on its way. Start with one deployment. Run the five decisions with its owner before you approve more spend.”
- How your details are used: to email you the guide and reply to you about it.
Or keep the screen result from Slot 7. It is yours either way. No email was needed for it, and none is needed to keep it.
If the verdict is unclear
If you ran the screen and the verdict is unclear, put the decision through the free consultation form. Tell the decision you are facing, what you have tried, and what evidence you have. One exchange is free. Replies go out in batches, usually within two business days. Details are used only to reply to the inquiry. They are not added to any mailing list.
Or run the AI-Ready L&D Function Assessment, the next diagnostic step the guide points to.
This is optional. It is not part of the guide, and it is not required to use it. The guide remains useful on its own if you never take this step.