AI is making it easier than ever to build software, products and polished concepts. Alex Osterwalder’s warning is that this speed can also make it easier to build the wrong thing, because a convincing product is not evidence that customers care.
In this episode of Curious by Nature, Gregory Gjini speaks with Alex Osterwalder, co-founder of Strategyzer and co-creator of the Business Model Canvas and Value Proposition Canvas. The conversation connects more than two decades of Alex’s work on business models and innovation with questions that B2B SaaS and technology leaders are facing now: ICP, differentiation, AI disruption, efficient growth, experimentation and how to turn strategy into better decisions.
The full episode of Curious by Nature with Alex Osterwalder is available on YouTube, Spotify and Apple Podcasts.
Lytt til episoden i din favorittapp
TL;DR
Alex Osterwalder argues that AI does not change the fundamental questions of business: who the customer is, how a company creates value and how it captures value. Faster product development makes evidence and customer understanding more important because teams can now build polished products before proving demand. He also explains why ICP should emerge from evidence, why business models can be more defensible than products, and why companies need systems for testing and killing weak ideas.
About Alex Osterwalder
Alex Osterwalder is co-founder of Strategyzer and one of the creators behind the Business Model Canvas and Value Proposition Canvas, tools that have become widely used across entrepreneurship, corporate strategy, innovation, product development and consulting.
His work with Yves Pigneur began in academia. Alex’s doctoral research explored how business models could be made explicit and designed more like an architect designs a building. That work eventually developed into the Business Model Canvas, books, software, training and, more recently, Strategyzer’s Playbooks.
During the conversation, Alex says that Strategyzer can see millions of downloads of its canvases, while the true reach is impossible to measure because the tools have spread through companies, classrooms, consulting projects and management teams around the world.
Sign up and receive new Episodes and insight in your inbox
Key takeaways
- Strategy should be treated as a collection of hypotheses whenever important variables remain uncertain.
- Business Model Canvas makes assumptions visible, but evidence determines whether they are true.
- Better business models can create stronger competitive protection than superior products alone.
- B2B companies should map the buying ecosystem before deciding which personas and accounts matter most.
- ICP should become sharper through real customer and sales evidence.
- Customers care about value before they care about differentiation.
- AI dramatically accelerates building, but does not replace customer discovery or experimentation.
- Leadership teams should become better at retiring projects that fail to generate evidence.
What this episode is really about
The familiar Strategyzer canvases are the starting point, but they are not the main subject of this conversation.
“Better business models will always kill better products.”
Alex Osterwalder
The more interesting question is what happens after a strategy has been put on a canvas.
Alex argues that a strategy involving significant uncertainty is still a collection of assumptions until there is evidence behind it. A company can describe a compelling future, build a detailed business case and create a polished product, while still being fundamentally wrong about the customer, their willingness to pay or the economics of acquiring them.
This distinction runs through the entire episode. Strategy is not simply about choosing an idea and executing harder. Leaders need to identify what must be true, test those assumptions and change direction when the evidence does not support them.
That same logic connects the discussion about business models, ICP, AI and Strategyzer’s evolution toward Playbooks.
“Ideas are everywhere, they’re free. What we need to do is make evidence-based decisions.”
Alex Osterwalder
Why this episode is relevant now
AI has dramatically lowered the effort required to create software, prototypes, messaging and content. Alex sees enormous potential in that development, but also a new risk.
Teams can now produce software that looks finished before they have proven that the underlying customer problem deserves to be solved. A polished interface can create a false sense of product-market fit.
Alex refers to an observation from Steve Blank’s Lean LaunchPad work: teams are becoming extremely good at building, while the harder discipline remains learning whether they are building the right thing.
For software companies, this creates a second challenge. If relatively simple tools can increasingly be recreated with AI, the product alone may become a weaker source of protection. Alex expects small tools solving narrow jobs to be particularly exposed and believes companies should explicitly assess the AI disruption risk of their business model.
The consequence is not that technology matters less operationally. It is that technology alone does not answer the fundamental business questions: Who is the customer? How do we create value? How do we capture value?
Who should listen to this episode?
- This conversation is particularly relevant for:
- Founders and CEOs building B2B SaaS or technology companies
- CMO, CRO and GTM leaders working with positioning, ICP and growth
- Product and innovation leaders responsible for new products or business models
- Boards, Private Equity and Venture Capital investors evaluating future growth
- Sales and marketing teams working with complex B2B buying processes
- Strategy consultants and internal transformation teams
- Leaders deciding where AI creates opportunity and where it creates disruption risk
– Flere episoder om B2B SaaS, salg og kommersiell skalering
The most important insights from S2 – Episode 8
1. Strategy fails when assumptions are treated as facts
One of Alex’s central arguments is that companies frequently confuse a coherent strategy with a validated strategy.
When a company is improving an existing product for customers it already understands, many variables are known. The situation changes when the company enters a new market, targets a new customer, introduces a different revenue model or builds something genuinely new.
At that point, a Business Model Canvas or strategy document describes what the company believes will work. It does not prove that it will.
Alex’s preferred question is therefore simple:
What needs to be true for this strategy to work?
From there, leaders can identify assumptions about customers, demand, willingness to pay, acquisition and economics, then collect evidence before committing increasing amounts of capital.
“We overthink and we underexperiment.”
Alex Osterwalder
That is a useful correction for leadership teams that spend months refining plans while doing relatively little to reduce the uncertainty behind them.
2. Better business models can beat better products
One of the strongest statements in the episode is Alex’s argument that companies often focus too heavily on product and technology while underestimating the competitive power of the business model.
“Better business models will always kill better products.”
Alex Osterwalder
He uses GoPro, WHOOP, Oura and Apple to illustrate different aspects of the argument.
A successful product can still be fragile if competitors can reproduce it and the company lacks strong recurring revenues, switching costs, ecosystem advantages or other forms of protection. Alex points to Apple’s ecosystem as an example of how several reinforcing business model advantages can create a much stronger position than product features alone.
For SaaS and technology leaders, this shifts the question from “How do we make the product better?” toward “How strong is the system through which we create, deliver and capture value?”
3. ICP in B2B starts before the ICP
The ICP discussion produces one of the most useful B2B lessons in the episode. Alex pushes the conversation one step back. Before defining an Ideal Customer Profile, companies need to understand the customer ecosystem surrounding a purchase.
Technology and product teams often validate their solution with users because those users are accessible and understand the product. But the user may have little influence over whether the company buys.
In a complex B2B sale, the relevant stakeholders might include a user, economic buyer, CEO, CIO, plant manager or channel partner. The right combination changes with the context.
Strategyzer’s approach is to map the relevant customers, users and stakeholders first, then identify the small number that truly matter for the purchase.
Only after companies start selling and gathering evidence can a stronger ICP emerge. Alex gives the example of contextual signals such as a new CEO making a company more likely to buy.
The practical implication is important: an ICP should not simply be invented in a workshop and treated as permanent. It should improve as the company accumulates evidence.
4. Customers care about value before they care about differentiation
Many marketing and strategy teams start by asking how the company is different from competitors.
Alex challenges that order in this episode.
From the customer’s perspective, the first question is whether the company creates meaningful value. Differentiation becomes relevant once the customer believes the solution is valuable and needs to decide between alternatives.
This sounds simple, but it changes how companies approach positioning.
A business may proudly claim that its solution is faster, more advanced or powered by better technology. None of those differences matter if the customer does not value the resulting benefit.
The correct sequence is therefore to establish customer value first, then demonstrate why the company’s approach delivers that value better than the alternatives.
5. AI makes building easier, but learning remains hard
Alex’s Fireflies.ai example captures this particularly well.
Before automating its meeting-note product, the founders reportedly joined customer meetings themselves, stayed silent, took notes and returned summaries as if the service were already automated. They learned from the customer experience before investing heavily in building the technology.
AI now allows founders to reverse that sequence because building has become so fast.
A team can create a polished application in days, show it to people and feel that substantial progress has been made. Alex’s concern is that the visible quality of the product can hide the absence of evidence.
This creates a paradox: as the cost of building falls, the discipline of not building too early may become even more valuable.
6. The best companies build systems for exploring, testing and killing ideas
Running an existing business and inventing its future require different operating systems.
Alex describes this through the familiar distinction between exploit and explore. Exploit means operating and improving the business that already exists. Explore means testing products, markets and business models where important variables remain unknown.
The problem is structural he says. Existing businesses have targets, customers, budgets and operational urgency. New opportunities are smaller, uncertain and initially less attractive. Without deliberate protection, the existing business tends to kill the new one.
This becomes especially practical when Gregory gives Alex a fictional PE or VC-backed Nordic SaaS company with slowing growth, unclear ICP, an undifferentiated value proposition and pressure to respond to AI.
Alex’s first recommendation is not a large transformation program. He would start by assessing the company’s AI disruption risk and the strengths and weaknesses of its existing business model. From there, leadership can generate alternatives, test several ideas and progressively kill the ones that fail to produce evidence.
For a smaller company, Alex suggests exploring several ideas rather than placing everything on one bet. Larger companies should manage a much broader portfolio and become comfortable retiring a substantial proportion of projects as evidence develops.
The point is not to celebrate failure. It is to prevent weak ideas from consuming resources simply because nobody wants to stop them.Dette gir en mer risikoreduserende måte å skalere på. Først når selskapet ser signaler på repeterbarhet, kan man investere tyngre.
From Business Model Canvas to Strategyzer Playbooks
The episode also explains an important evolution in Strategyzer itself.
Books made the frameworks accessible, but Alex says books are limited when the goal is to change how people actually work. Strategyzer experimented with apps, software, e-learning and workshops before arriving at what the company now calls Playbooks.
A Playbook breaks a business problem into a sequence of actions. Instead of watching a course about value propositions, for example, a team learns a step and immediately performs the work before moving forward.
Alex compares this with learning tennis. Watching videos about how Roger Federer holds a racket does not make somebody play like Federer. Learning improves when instruction and practice happen together.
Strategyzer currently applies this model to areas including value propositions, business models, testing, team alignment and AI-assisted experimentation. The company is also working toward making its methods accessible to teams, individual practitioners, independent consultants and coaches, rather than reserving the knowledge for large consulting engagements.
What this means in practice and per function
For CEOs and leadership teams
Separate the business you are running from the business you may need to build next. Both matter, and neither should automatically dominate the other.
For significant strategic bets, ask teams to identify their most important assumptions and show evidence behind them. A confident presentation is not evidence, and neither is a detailed spreadsheet.
For marketing, sales and GTM leaders
Revisit ICP as an evolving commercial hypothesis rather than a static segmentation document.
Map the buying ecosystem, identify which stakeholders actually determine the purchase, and build value propositions around what each critical stakeholder needs to achieve.
Do not start differentiation work with competitors. Start with customer value.
For product and innovation teams
Do not allow cheaper and faster AI-enabled development to remove customer discovery from the process.
The ability to build something quickly should make experimentation faster, not make experimentation unnecessary.
For boards, PE and VC investors
Efficiency and exploration are not opposites.
A company can improve the economics of its existing operation while still investing deliberately in future business models. The board’s influence starts with the type of CEO it appoints and the expectations it sets around experimentation, evidence and resource allocation.
The relevant question is not simply whether management has an AI initiative. It is whether leadership understands where AI threatens the current business model and where it creates new ways to create and capture value.
Key takeaways
- Strategy should be treated as a collection of hypotheses whenever important variables remain uncertain.
- Business Model Canvas makes assumptions visible, but evidence determines whether they are true.
- Better business models can create stronger competitive protection than superior products alone.
- B2B companies should map the buying ecosystem before deciding which personas and accounts matter most.
- ICP should become sharper through real customer and sales evidence.
- Customers care about value before they care about differentiation.
- AI dramatically accelerates building, but does not replace customer discovery or experimentation.
- Leadership teams should become better at retiring projects that fail to generate evidence.
FAQ
Who is Alex Osterwalder?
Alex Osterwalder is co-founder of Strategyzer and co-creator of the Business Model Canvas and Value Proposition Canvas. His work focuses on business model design, value propositions, innovation, experimentation and turning strategy into practical tools and processes.
What does Alex Osterwalder say about AI and software companies?
Osterwalder argues that AI makes software faster and cheaper to build, but this can cause teams to build before validating the customer problem. He expects narrow software tools solving relatively small jobs to face greater disruption risk and recommends assessing how AI could affect the entire business model.
Why does Alex Osterwalder say better business models beat better products?
His argument is that product quality alone may be easy to copy or disrupt. Stronger business models can add recurring revenue, ecosystems, switching costs and other forms of protection that make a company more resilient even when competitors can build comparable products.
How should B2B companies define an Ideal Customer Profile?
According to Osterwalder, companies should first map the customers, users and stakeholders involved in the purchase. They should prioritize the people who matter most to the decision, then allow the ICP to become more precise as real sales and customer evidence reveal the contexts in which companies are most likely to buy.
What is the difference between customer value and differentiation?
Customer value comes first. A buyer initially cares whether a product solves an important problem or creates a meaningful benefit. Differentiation matters when the customer compares that value with competing alternatives. Being different is therefore only useful when the difference matters to the customer.
What is the difference between explore and exploit?
Exploit means running and improving an existing business where customers, products and economics are relatively well understood. Explore means testing new products, customers or business models where uncertainty is high. Companies need both, but the existing business often needs to be prevented from automatically killing new opportunities.
What are Strategyzer Playbooks?
Strategyzer Playbooks are structured, step-by-step workflows that help teams apply methods around value propositions, business models, experimentation and strategy. Instead of only learning a framework, users perform the work during the process and produce concrete outputs that can be tested and improved.
Key terms and terminology mentioned in this episode that you might wonder about
Business Model Canvas
A visual tool developed by Alexander Osterwalder and Yves Pigneur for describing how an organization creates, delivers and captures value.
Value Proposition Canvas
A Strategyzer tool for connecting customer jobs, pains and gains with the products, services and benefits a company offers.
Ideal Customer Profile (ICP)
A description of the type and context of customer most likely to create a strong commercial fit. In Alex’s approach, it should be informed by evidence rather than firmographics alone.
Explore
The process of searching for and testing new products, markets, value propositions and business models under conditions of uncertainty.
Exploit
Running, optimizing and scaling a business model that is already understood.
Evidence-based strategy
Making investment and scaling decisions based on observed customer behavior and experiments rather than opinions, forecasts or internal conviction alone.
AI disruption risk
The degree to which AI could weaken important elements of an existing business model or allow competitors and customers to replace parts of the value currently provided.
Strategyzer Playbook
A structured workflow that combines instruction with practical activities so teams can apply Strategyzer methods to a specific business problem.
Comments (0)