California Management Review
California Management Review is a premier professional management journal for practitioners published at UC Berkeley Haas School of Business.
by Vijay Govindarajan, Daniel J. Finkenstadt, and Tojin Thomas Eapen
Image Credit | Artelio
In today’s volatile world, conventional thinking is a liability. To stay ahead, businesses must flip expectations and embrace bold, unconventional strategies. Reverse thinking is one such powerful mental model. It’s not new. Legendary 19th century mathematician Carl Jacobi urged, “Invert, always invert.” Warren Buffett’s best-known advice? “Be fearful when others are greedy, and greedy when others are fearful.” The logic is timeless: doing the opposite forces deeper insight.
Sara L. Beckman. “To Frame or Reframe: Where Might Design Thinking Research Go Next?” California Management Review, 62/2 (2020): 144-162.
Eric Knight, Jarryd Daymond, and Sotirios Paroutis. “Design-Led Strategy: How To Bring Design Thinking Into The Art of Strategic Management.” California Management Review, 62/2 (2020): 30-52.
Reverse thinking serves as a powerful nexus between critical and creative thinking. While critical thinking ensures that ideas are systematically evaluated, creative thinking fuels the generation of novel solutions. Businesses can leverage this synergy by first deconstructing existing paradigms through critical analysis and identifying foundational assumptions that may limit innovation in three ways.
Reverse thinking can be used in three ways. First, one can use assumption reversal as a search principle to uncover hidden business opportunities, such as new product attributes or business models. Second, role reversal can be used as a testing principle to find opposing viewpoints and uncover weak points. Finally, it can be employed for tactical reversal, where counterintuitive actions can yield favorable results.
AI can turbocharge this process. The application of reverse thinking in business has often been limited by the constraints of human cognition. Large language models (LLMs) can help identify hidden biases, generate contrarian ideas, and test unconventional paths quickly and at scale. Firms can leverage generative AI for idea inversion, highly specialized narrow AI for risk detection, graph-based retrieval (GraphRAG) to surface non-obvious connections, and AI agents to simulate stakeholder reactions, all at once, enabling structured, reverse thinking that challenges assumptions and reveals strategic white space.
By flipping assumptions about attributes, reverse thinking expands the idea space and fights groupthink. This has application in product development and identification of new business models.
Recently, Japanese cosmetics giant Shiseido flipped beauty product development on its head using reverse thinking. Instead of adding skincare benefits to a foundation, a well-trodden path, their R&D team started with a serum and embedded foundation properties into it. The result? A game-changing hybrid: the “foundation serum,” delivering breakthrough benefits in both make-up and skincare. Launched in 2023–24, it became a big hit – achieving the #1 market share in Japanese department-store foundations and attracting a wave of younger customers to the brand.
Flipping the script on conventional thinking can unlock breakthrough business models. India has over 100 million people in need of eye care, but fewer than 20,000 ophthalmologists. The industry’s dominant assumptions? Eye screening devices must be expensive (>$50,000), bulky, operated by trained specialists, and require eye dilation. Forus Health turned these assumptions on their head. Instead of accepting cost, complexity, and clinical environments as immovable constraints, they asked: What if we designed for affordability, portability, and ease of use — no specialists or dilation required? That reversal led to a $10,000, lightweight, portable scanner usable in non-clinical settings, screening for cataracts, diabetic retinopathy, glaucoma, and more. Forus didn’t just tweak the model, they rebuilt it around flipped assumptions. The result? A profitable company, active in 60 countries, improving vision for 16+ million people worldwide.
AI helps to uncover hidden opportunities by flipping assumptions and challenging business norms. Generative models like large language models (LLMs) can reframe standard innovation questions, e.g., “What do customers want?” becomes “What do they not want, and why might that be valuable?” to spark unexpected ideas. Prompts such as “What if we removed our best feature?” or “How could our biggest flaw become our core differentiator?” help teams explore overlooked possibilities. To expand the innovation space, firms can integrate narrow AI systems and graph-based methods with LLMs. Narrow AI can surface anomalies or counterintuitive patterns. For example, customers who churn due to excessive value, while graph-based AI reveals hidden relationships between user behaviors, products, or market dynamics. These insights help identify tensions or edge cases that would otherwise go unnoticed.
When used in combination, these AI tools form a powerful discovery engine. Graph algorithms highlight structural contradictions; narrow AI models validate and prioritize areas of interest; and LLMs synthesize these into provocative, scenario-based prompts for exploration. For instance, anomaly clusters surfaced in usage graphs might inspire the question, “What would a product look like if it only served customers who avoid automation?”
Adopt the devil’s advocate mindset. In 1587, Pope Sixtus V institutionalized this concept to rigorously vet candidates for sainthood. Today, it helps companies to reveal weak spots, pressure-test ideas, and avoid blind optimism.
Netflix’s “Chaos Monkey” is a prime example of reverse thinking as a testing principle: instead of merely striving to keep systems running smoothly, it deliberately injects failure into live production environments. Developed by Netflix’s engineering team, the tool randomly disables servers, forcing the rest of the infrastructure to adapt on the fly. In effect, Netflix built controlled chaos into its daily operations, compelling its systems to be designed with high resilience from the start. A 2020 survey by JD Power found that it was considered the most reliable streaming platform among U.S. users. Since then, the company’s revenue has grown from $25 billion USD in 2020 to $39 billion USD in 2024.
Large Language Models (LLMs) such as GPT4o, Claude and Gemini can simulate opposing perspectives to pressure-test unconventional ideas and expose blind spots. By prompting these models to act as skeptical investors, risk-averse regulators, or aggressive competitors, teams can quickly generate critiques that challenge groupthink and reveal potential pitfalls. For example, before removing a key product feature to simplify the offering, an LLM could forecast customer outrage, legal risk, or unintended brand consequences. These “virtual critics” serve as low-risk simulation tools that help stress-test ideas early. Prompts like “Simulate a competitor’s counter-move” or “What would a customer complaint look like six months after launch?” push thinking beyond the optimistic case and build strategic resilience. Entrepreneurs can test their business ideas by pitching to a shrewd AI investor, as we demonstrate in our custom GPT, Lair of the Lion: bit.ly/lairofthelion.
This process becomes even more powerful when LLMs are combined with narrow AI, graph-based analysis, and agent-based orchestration. Narrow AI models can quantify risks (e.g., churn probabilities or sentiment shifts), while graph analytics surface latent relationships among features, users, or complaints, helping teams understand who will be affected and how. Multi-agent systems can coordinate these tools, assigning roles like “customer advocate,” “compliance officer,” or “disruptive competitor” to AI agents that interact and debate scenarios in real time. Together, this ensemble can uncover second- and third-order effects, identify narrative inconsistencies, and strengthen decisions before they reach the market, shifting innovation from reactive to anticipatory.
Sometimes, doing the opposite works. Burn bridges to boost resolve. Raise prices to build the perception of high quality and increase demand. Fake a retreat to win a war. Xiang Yu did it in 210 BC when he destroyed his own ships to eliminate retreat as an option, and his troops fought harder and won.
Some companies have successfully employed reverse marketing approaches, deliberately contradicting conventional messaging to attract attention and specific audiences. A celebrated example of reverse thinking using counterintuitive tactics comes from Patagonia’s iconic “Don’t Buy This Jacket” campaign, during the 2011 Thanksgiving season. Following the campaign, the company experienced a remarkable 30% surge in revenue, reaching $543 million in 2012, followed by an additional 5% growth in 2013. Today, the company has over USD 1.5 billion in revenue.
Anduril Industries recently flipped conventional hiring logic on its head. In February 2025, the defense tech startup founded by Oculus VR creator Palmer Luckey launched a jarringly counterintuitive hiring campaign: “Don’t Work at Anduril.” Graffiti-style versions of the slogan appeared in tech hotspots like Boston, Atlanta, and Seattle, spray-painted right over their typical recruitment ads. It looked like sabotage. It was brilliant marketing.
The unexpected twist went viral, sparking buzz on LinkedIn and beyond. “Anduril is not for everyone. That’s the point,” said Jeff Miller, VP of Marketing. The message filtered out anyone who wasn’t mission-driven and attracted the right kind of bold talent to fill over 700 open roles. By flipping the script on traditional hiring, Anduril sparked conversation and sharpened its brand edge. Based on Google Trends data from the United States, this campaign resulted in by far the largest increase in online searches for the company since its founding in 2017, surpassing several established competitors. According to Miller’s post on Linkedin, job applications at the company went up by 30% following the campaign. Anduril’s YouTube video on the campaign became their best-performing video upload, viewed over 3.6 million times in the next two months.
AI can help companies design and refine bold, counterintuitive actions that command attention and generate disproportionate impact. By prompting LLMs to invert conventional tactics, such as replacing mass promotions with scarcity-driven limited drops or using anti-advertising to provoke curiosity, teams can explore strategies that cut through noise while aligning with brand purpose. Prompts like “Suggest a bold opposite action that fits our values” or “Write a warning message that paradoxically attracts the right audience” allow firms to surface ideas that are both disruptive and on-brand. LLMs can also adapt tone and framing to specific segments, ensuring provocative messaging doesn’t veer off-message.
When paired with narrow AI and predictive modeling, this becomes a strategic experimentation engine. Sentiment analysis, behavioral prediction, and A/B simulation tools can forecast how different audiences will react to bold positioning, while graph-based AI can identify influencer clusters or subcultures likely to amplify the message. Combined in agent-based systems, one agent might generate unconventional ideas, another might simulate audience reception, and a third might optimize channel targeting, all in a closed loop. This systemized approach helps de-risk bold moves by previewing their emotional, reputational, and commercial ripple effects before execution.
Reverse thinking isn’t about being contrarian for the sake of it. It’s about outmaneuvering the obvious. And when powered by AI, its potential is exponential. Those who dare to flip the script, who question everything, invert expectations, and use AI to imagine radical alternatives, won’t just survive disruption. They’ll define what comes next.
