When an AI initiative dies, the machine takes the blame. It's almost always the wrong suspect.
The bug isn't in the technology. It's in us. A working diagnostic for the chain of management decisions made long before anyone trains a model.
After ordering: claim your free bonus stack · or join the launch team for early access After ordering: claim your free bonus stack · or join the launch team
Available in hardcover and ebook · ISBN 978-1-394-45493-8
Already read it? Help the next reader find it.
A short review on Amazon or Goodreads in launch week is the single biggest favor you can do this book.
"Time to update?"
The message was three words long: no preamble, no hope-you-are-well. The investors, board members, and senior executives who'd read Digital Transformation Payday were watching companies repeat the same mistakes with AI — and they wanted the next book.
PREVAIL is that book. Eight unasked questions. One pattern, repeating. Written for the boardroom, not the seminar.
Read the Preface and Chapter 1: Eight Hard Questions
The full opening of the book: the three-word message that started it, the eight questions that frame the chapters, and the methodology in plain language. No email required. About 25 minutes.
PDF available shortly — the sample is with the publisher.
What executives are saying.
"Grounded in research and focused on real-world decision-making, this book is a valuable resource for executives, boards, and business leaders seeking to translate AI ambition into measurable impact."
"Bottke cuts through the GenAI hype to address the fundamental question every board and finance leader faces: how to bridge technological ambition with measurable bottom-line value. An essential, data-driven framework for any executive accountable for turning AI investments into real economic performance."
"PREVAIL is an essential read for executives seeking real AI outcomes rather than technology experimentation. The book’s strength lies in connecting business value, foundational readiness, and practical guidance, while emphasizing that AI demands different skills, governance, and expectations than traditional software development."
"I run a company that builds AI into its software, which makes me exactly the kind of reader this book is written to challenge. Bottke's point is hard to escape: the technology rarely fails on its own. The decisions around it do, and in the same order every time. PREVAIL turns AI from a technology question into a leadership question."
"Bottke has spent his career in boardrooms where extremely difficult questions about technology are asked. He is known for his ability to provide simple and clear answers to immensely complex problems. Useful, grounded, and backed by evidence, PREVAIL lives up to this reputation."
"Bottke does something rare in business writing: he makes a provocative argument about AI management and shows his work. PREVAIL combines rigorous evidence with a practical framework to help leaders distinguish AI activity from genuine business value."
"Bottke's PREVAIL provides an essential framework for any decision-maker navigating the complexities of any bought or built AI deployment."
"In PREVAIL Bottke captures extremely well the challenges that adoption of AI brings to organizations and how to understand the relationship between business strategy and AI. It is a very practical guide for leaders to ask the right questions to ensure success with AI."
"Concrete, and unsentimental. Grounded in hard analysis, it asks the tough questions others avoid. Leaders who must decide when and how to invest will love it."
"PREVAIL is the book every R&D team wants their CEO to read."
"PREVAIL is the black-box recording of your AI transformation crash. Only you get to hear it before it happens."
"Most AI books are about what the technology can do. In PREVAIL, Bottke focuses on whether organizations are ready to put it to work and create value. He follows the management decisions that often lead from an impressive demo to an expensive disappointment and equips managers with the questions to challenge those decisions while there is still time to change course."
"A refreshingly rational and detailed look at one of the most urgent questions in conference and board rooms today. Bottke thoughtfully lays out, analyses, and provides solutions for systemic blind spots faced by most leadership teams."
"AI in business demands discipline, not hype. This book gives executives a clear, evidence-based framework to define problems, test rigorously, and scale AI with accountability — turning pilots into measurable impact."
"Some of AI's greatest risks come from the human choices surrounding it. Drawing on evidence and experience, Bottke offers insight into those choices and their consequences. PREVAIL is a timely and practical guide for leaders seeking to approach AI with greater clarity and understanding."
"For decades I've seen companies struggle with innovation for the same reasons they're now struggling with AI. Bottke masterfully names that pattern and turns it into practical tools courageous leaders can act on immediately."
"Pragmatic and useful for AI and agentic business success. And to discover where this investment actually adds to market value."
17 endorsements · scroll or use the arrows
Where did your AI program stall?
P-R-E-V-A-I-L, read from the ground up. Six floors and one elevator. The ground floor is where leaders reach for the tool before they've named the problem — and the uncomfortable finding is that the damage concentrates there and compounds as management builds on those decisions. The elevator is L (agents). It rides every floor at machine speed. The market sits outside the building and prices only what shows up in the results. Eight questions. The book takes them in order.
How I know.
18,130 publicly available company filings from 978 listed companies, 2021 through early 2026. The pipeline surfaced 3,518 paragraphs that talk about AI. After filtering, two-stage LLM classification, and line-by-line human review, 48 filings from 29 companies remained — validated AI breakdown signals, written by the firms themselves. Here is how I got there.
Two-stage LLM classification, then the part no one automates: every flagged paragraph read by hand over a three-week stretch. At the confirmed and probable tiers: 100 percent binary precision. Zero false positives. Every filing the pipeline flagged as a high-confidence AI breakdown was, in fact, an AI breakdown when a human read the source text independently. It was miserable work.
Methodology one-pager available shortly — the full method is in the book's appendix.
Filing language is flat, but precise in ways that reward close reading. Most disclosures never mention AI breakdowns in those terms — so the pipeline reads for what the language does. Three patterns from the real world:
The program is working. The filings have moved on — the language has integrated into segment results, attached to a function, a metric, a cost line.
Enthusiasm decaying quarter by quarter. Mentions shrinking. No announcement of breakdown — but no continued investment either.
Language appeared for three or more quarters, then vanished. No explanation, no follow-through, no disclosure of what happened to the initiative.
Signal patterns as defined in the book — no company's filing is quoted or reproduced.
Seven questions for your next AI review.
Not board questions — the ones the people who build and run your AI should answer cold. One per floor, pulled from the book's 42-probe self-assessment. Put the AI lead, the data engineer, and the CFO in a room and ask. The ones that get a pause are the floor where the breakdown is already running. And the one question that predicts success better than anything: what problem does this solve?
Tear-out card · one per floor
Score where you're weakest, then start there. Problems on lower floors cascade upward.
- StrategyCan you state, in one sentence, the business metric your most important AI initiative is meant to move?
- DataIs there a named owner — a person, not a committee — for every dataset your AI relies on?
- ProcessIf a model failed at 2 a.m. Saturday, is there a tested rollback to the pre-AI process?
- OrganizationDo you measure AI adoption by the decisions it changed, or just by who has access?
- MeasurementCan you state the annual cost to operate your AI — not to build it, to run it?
- LifecycleIs there a single list of every AI model you have in production right now?
- AgentsIs each agent's authority written down — what it may do alone, what needs sign-off, what it must never do?
Seen enough? Lock in your copy — and the bonus stack.Seen enough? Grab your copy — and the bonus stack.
Pre-Order on Amazon.com →Buy on Amazon.com →Tim Bottke
For journalists, producers, and event organizers.
Everything you need to cover the book or book the author: press release, author bio, book fact sheet, a quotable Q&A, and a methodology note for journalists. High-resolution headshot and cover art on request.
Download the press kit
Press release, author biography (short and full), fact sheet with ISBN and publication data, and an author Q&A cleared for quotation.
Get in touch
Interview requests, review copies, podcast bookings, and keynote inquiries. Replies within one business day during the launch window.
Help me hit the lists. Here's the deal.
Pre-order, upload your receipt, and you read PREVAIL months before it’s out. In launch week I’ll ask for two things: an honest review and one share on LinkedIn. Nothing you get depends on whether you review the book, or on what you write.
What you get
- Advance reading copy via secure publisher linkSent by email once your receipt is in
- The full bonus stack, two weeks earlyLate December
- Private Q&A video: your questions, my answersLaunch week
- Your personal launch-team cardAfter launch
What I ask
- Pre-order now (any retailer)Upload the receipt with the form, or later
- Post an honest reviewAmazon or Goodreads, in launch week
- Share the book on LinkedInReady-to-post draft available
Buy the book, get the working materials.
Buying the book is step one. Everything below follows from it — the bonus stack for individual readers, tiered programs for teams and boards.
Three steps. Launch-day delivery.
- 1Pre-order from any retailer, anywhere in the world
- 2 — a photo or PDF from any retailer works
- 3Stack arrives on launch day, Jan 14 (after launch: within 24 hours)
What's in the stack
- 42-probe diagnostic (Appendix A worksheet)PDF
- Board-ready slide deck (16 slides)PPTX
- Author webinar, recordedVIDEO
Receipts dated through Feb 14, 2027 qualify — already ordered? Still counts.
Tiered pricing for groups, offsites & clients.
- 1Order via Porchlight — the bulk specialist, with quantity discounts and sales that count toward the lists. (no store account? request a Wiley quote in step 2)
- 2 — tier benefits attach to the registration, wherever you bought
- 3On confirmation, the author schedules your tier benefits directly
Estimated quantity
You prevail because your management did —
not the machine.
It breaks the same way every time.
Fix the sequence, not the technology.
The questioning is your job. It always was.
+ Free bonus stack with every order —
how to claim