The StackedCFO Series · Book 4 · Casper Zhao

The Deal
Is AlreadyWrong.

How AI Is Rewriting Due Diligence, QoE Work, and the Economics of Every LBO You Touch

By Casper Zhao, CPA · StackedCFO LLC
Get the Book Read Quotes
The Deal Is Already Wrong — Book Cover
The addback that passed three prior diligence processes just got flagged in 40 seconds The QoE firm isn't selling insight. It's selling defensibility. AI can now manufacture both The deal is already wrong if you haven't started paying attention Software diligence used to be sampling. Now it is comprehensive. The difference shows up in the returns The loan tape knows things management has not yet disclosed The model was right. The inputs were wrong. AI is fixing the inputs The addback that passed three prior diligence processes just got flagged in 40 seconds The QoE firm isn't selling insight. It's selling defensibility. AI can now manufacture both The deal is already wrong if you haven't started paying attention Software diligence used to be sampling. Now it is comprehensive. The difference shows up in the returns The loan tape knows things management has not yet disclosed The model was right. The inputs were wrong. AI is fixing the inputs
Who This Book Is For

Six people in the room.
All of them need this.

QoE Practitioners
Your first draft now takes hours, not weeks. Do you know what to do with that?
AI has compressed the production layer of your engagement. This book explains what expands in its place — and why the practitioner who doesn't make that shift becomes the commodity version of the deliverable.
Buy-Side Deal Teams
The data room has no more secrets — for the buyer who has the right tools.
Cohort retention from primary data. Working capital from five years of monthly balance sheets. Contract populations reviewed in full rather than sampled. This book maps exactly what AI enables and where human judgment still wins the deal.
Sell-Side CFOs
The buyer already knows what your data room says. Did you know what it said first?
The management interview you're preparing for is now a room where the other side has already cross-referenced every claim in your presentation against the underlying data. This book tells you how to prepare for that reality rather than be surprised by it.
Advisory Firm Leaders
Your fee is being justified by production hours that AI just made cheaper to replace.
The leverage model that built the QoE industry is under structural pressure. This book walks through what the winning advisory firm looks like in five years and what it stops doing today to get there.
Lenders and Credit Teams
The covenant you set at 5.5x would have tripped twice in the last three years. AI would have told you that before signing.
Loan tape analysis. Covenant design from historical compliance modeling. Independent verification of the QoE you received. This book covers how lenders are building underwriting advantages that will show up in credit performance over the cycle.
Junior Practitioners
You're building models faster. Are you learning the judgment that makes those models right?
The development path for a diligence professional has changed. This book explains what skills are being automated, what expands, and how the practitioners who figure this out early will have a significant structural advantage over the ones who don't.
10 Reasons to Open This Book

The sentences that
start the argument.

01
There is a number near the top of almost every Quality of Earnings engagement letter. That number has been remarkably stable for a long time. That stability is about to end.
02
The QoE firm charging $400K for a six-week engagement is now competing with a two-person team and an AI stack that delivers a normalized EBITDA bridge in 10 days. The buy-side associate building the model manually is sitting next to someone who prompts their way to the same output in an afternoon.
03
The addback schedule is a negotiating position. Every dollar on it, if accepted, increases the purchase price at any given multiple. AI just gave the buyer the tool to evaluate every line before the first management call.
04
A management team claiming that a specific expense is non-recurring is making an empirical claim about the past. The general ledger contains the evidence. AI reads five years of transaction detail in seconds.
05
The buy-side AI analysis runs a revenue-weighted cohort retention calculation. It finds 78% retention by revenue. The management presentation said 92%. Both figures are technically accurate. Only one of them prices the deal correctly.
06
A software company with 105% blended NRR might have 20% of its customers expanding aggressively while 40% are contracting. The blended figure looks healthy. The distribution reveals a leaky bucket. AI sees the distribution in the time it used to take to set up the spreadsheet.
07
The healthcare services business has less room to obscure unfavorable patterns than it has ever had. The claims database is the evidence. The AI analysis is comprehensive. The expert interpretation that determines what the evidence means is not optional.
08
The advisory firm that cannot say, specifically, what its senior judgment added to an engagement — beyond what an AI-assisted analysis would have produced — is pricing its fee on faith. The market is losing that faith.
09
The lender who closes a credit and then faces simultaneous renewal negotiations with customers representing 40% of ARR in year one is managing a known risk they didn't know was concentrated. AI would have built the renewal calendar before the credit agreement was signed.
10
The deal table in 2030 still has people at it. They got there by understanding, earlier than everyone else, what the table was becoming. That understanding is available now. The question is who is building it.
10 Reasons This Book Is Different

What you won't find
anywhere else.

01
It names the specific failure modes, not just the capabilities
Every chapter that covers what AI can do also covers where it fails silently. The confident wrong answer is more dangerous than the visibly incomplete one. This book tells you which tasks produce which.
02
The sector exposure map exists nowhere else
Nine sectors. Specific AI diligence exposure percentages. Exactly what drives the ceiling in each one — whether it's regulatory interpretation, physical inspection, or domain expertise that structured data cannot replace. Software at 25%. Financial services at 8%. The differences matter.
03
It's written from the advisor's seat, not the technology vendor's
Casper Zhao ran 100+ QoE engagements at PwC, BDO, and CFGI. This book describes AI's impact on deal diligence the way someone who has done the work for 13 years describes it — specifically, honestly, with the failure modes included.
04
The pre-sale CFO checklist is a six-to-twelve-month action plan
Six categories. Thirty-two specific items. Everything a finance leader needs to do before a sale process launches to ensure the buyer's AI analysis doesn't surface things management should have found first. Most books describe the problem. This one gives you the protocol.
05
It covers both sides of every deal
Buy-side and sell-side. What each side can do with AI. What each side is now exposed to from the other. The advisor who understands the other side's analytical capability wins the room. This book is the other side's playbook.
06
The talent development problem is named and addressed
When AI produces the first draft, the associate who only reviews AI output develops different — and in some ways thinner — judgment than the one who built the analysis from scratch. This book explains why that matters for the next decade and what to do about it.
07
The covenant design chapter will change how lenders structure deals
Historical compliance modeling before the credit agreement is signed. The covenant set at 5.5x that would have been tripped twice in the trailing three years. Most lenders are setting headline numbers without this analysis. AI makes it possible on every deal.
08
The credibility reserve concept reframes every management interview
Management enters the diligence process with a finite amount of credibility. AI-assisted diligence depletes it faster because it surfaces more specific discrepancies sooner. The sell-side team that doesn't account for this is starting the interview already behind.
09
The independent practitioner opportunity is mapped in detail
The experienced QoE practitioner who left a large firm used to be capacity-constrained. AI has removed that constraint. This book explains exactly why the independent model that wasn't economically viable five years ago now is — and who is best positioned to capture it.
10
It ends with a direct practitioner's mandate, not a summary
Chapter 15 doesn't summarize. It asks you directly: what are you doing about this? Then it gives you the specific question for your specific role — buy-side, sell-side, lender, advisory firm leader — and explains why the answer you give now is more valuable than the one you give later.
10 Passages From the Book

The sentences
that hit different.

Ch. 1
The labor cost of financial analysis was the floor underneath the QoE fee structure. AI just removed the floor.
The Fee Model
Ch. 2
The EBITDA bridge AI builds is faster. It is not necessarily better. The gap between faster and better is where the practitioner's value lives now.
AI vs. Judgment
Ch. 3
The addback that passed three prior diligence processes just got flagged in 40 seconds. The model finds the dirty EBITDA. What the practitioner does with that finding is where the professional obligation lives.
Add-Back Scrutiny
Ch. 5
The summary statistic — an 88% retention rate — now comes with a distribution. A portfolio where 15 large customers retain at 99% and 200 smaller customers churn at 40% annually blends to the same number. The risk profile is completely different. The model built on the blended rate without understanding the distribution is priced incorrectly.
LBO Model Inputs
Ch. 6
The QoE firm isn't selling insight. It's selling defensibility. AI can now manufacture both. The QoE fee does not get to stay where it is. The market will not allow it.
Sell-Side Economics
Ch. 7
The sell-side QoE that anticipates the buyer's analysis wins the negotiation before the buyer arrives. The sell-side advisor who uses those tools on their own materials before the buyer does is not just being diligent. They are competing effectively.
Sell-Side Preparation
Ch. 8
Every number in the management presentation is now a commitment. The buyer has the tools to find out whether you kept it. The claim did not fail because it was wrong. It failed because it was underspecified in a world where the buyer can run their own specification in hours.
Management Presentations
Ch. 10
Healthcare diligence without regulatory expertise is just data. Data without regulatory context is just noise. The tools are powerful. The domain expertise is not optional.
Healthcare Services
Ch. 13
The advisory firm that cannot say what its senior judgment added — specifically — is pricing its fee on faith. The market is losing that faith. The question is not whether to change. It is whether to change on the firm's own timeline or on the market's.
Advisory Firms
Ch. 15
The deal table in 2030 still has people at it. They got there by understanding, earlier than everyone else, what the table was becoming. The seats at the head of it are being taken by the people who showed up early. That understanding is available now.
The Future
10 Ways to Say It

The deal table
is being rebuilt.

01
AI didn't replace the deal. It replaced the excuse for not knowing.
02
The addback that survived three diligence processes just failed its fourth — in 40 seconds.
03
You're not paying for the analysis anymore. You're paying for the judgment that validates it.
04
The data room has no more secrets. The question is who reads it first.
05
If your EBITDA can't survive a 40-second AI scan, it wasn't real EBITDA.
06
The floor is gone. Build on what remains: expertise, judgment, and speed.
07
Every QoE now has two versions — the one you submitted, and the one the buyer built before the management call.
08
The seat at the head of the deal table goes to whoever showed up most prepared. Prepare differently now.
09
AI reads the loan tape in the time it used to take to request it.
10
The deal is already wrong. This book is how you stop being the last to know.

Read it
before the
buyer does.

15 chapters. 5 parts. 9 sectors mapped. A prologue, three appendices, and a practitioner's mandate on the last page. Written by someone who ran 100+ QoE engagements and watched the floor disappear in real time.

Kindle on Amazon Paperback on Amazon
StackedCFO LLC · Boston, MA · casper@stackedcfo.com · stackedcfo.com