Chapter 014　What Is Innovation in the Age of AI?

of AI What is innovation? In most companies it is treated as a question about technology. The substance is not technology. It is a question about where new value comes from. AI can generate ideas in effectively unlimited quantity. At the work of enumerating combinations, human beings no longer compete with AI. So where does the scarcity in innovation move to? This chapter answers that one question and nothing else.

1 The question — why it arises now

The word innovation has a clear point of origin. Innovation was defined as new combinations: existing elements assembled in a form that did not exist before (Schumpeter, 1934). A new good, a new method of production, a new market, a new source of supply, a new form of organization. Innovation is not invention. It is the discovery of a combination. A century later, the definition has not aged. In the second half of the twentieth century, enterprises tried to manage it. Research and development functions were created, budgets were set, and stage-gate reviews were built. Innovation became a function of the organization. In the twenty-first century the limits of management began to be discussed. Try small, learn fast. Bring in outside knowledge. Build with the customer. The methods were refined and the cost of failure fell. Then AI arrived. What happened was not an update to methodology. The work of enumerating combinations was itself mechanized. Candidate materials. Templates for a business model. New price structures. New customer segments. New pairings of technologies. AI lays out a thousand of these in minutes. It also handles a wider search space than a person does. Human beings struggle to conceive outside their own experience. AI has no such constraint. The generation of new combinations, in Schumpeter’s sense, is therefore no longer a scarce capability. From here the question changes. Innovation used to ask how to produce good ideas. That is no longer the question. In a world with a surplus of ideas, what decides innovation? Only the executive can answer this. The answer does not lie on the side of technology. It lies on the side of choosing, investing, building to completion, and the will to carry something all the way into society.

2 Conventional answers and their limits

Three conventional answers about innovation circulate widely. Each is partly right. Each stops working in the Age of AI. The first conventional answer: “Innovation comes from a superior idea” This is the most persistent one. So companies try to collect ideas. They run internal calls for proposals, hold workshops, and build contact with outside founders. The premise carries a tacit assumption: ideas are scarce. That assumption has collapsed. Give generative AI the conditions and it produces a hundred new business proposals. It supplies the technology pairings, the entry routes, and the revenue structures alongside them. The quality is not low. Proposals at the level a planning team would take a week to produce arrive in tens of minutes. So when ideas multiply a hundredfold, does innovation multiply a hundredfold? It does not. The bottleneck has moved. Generation is solved. What remains is choosing, betting, building to completion, and getting society to accept the result. All four are questions of will. None is a question of information processing. AI therefore cannot solve them. In many organizations a surplus of ideas paralyzes decision-making instead. A meeting that can decide among three options cannot decide among a hundred. There is an asymmetry here. The cost of generation has fallen sharply. The cost of implementation has barely fallen. The plant still has to be built. The approval still has to be obtained. Customer habits still change only over time. The gap between what got cheap and what did not is the new bottleneck. The second conventional answer: “Innovation means discovering a market need” The second answer looks more practical. Listen to the customer. Study the market. Find the unmet demand. The method works in existing markets. It has one decisive limit. A market that does not yet exist cannot be surveyed. Before the smartphone, almost no consumer said they wanted a computer they could carry. Before electric vehicles spread, no survey showed the size of that market. Before generative AI spread, demand for paying a monthly fee to software that writes prose appeared in no market data. These markets were not discovered. They were created. Here is the paradox of the Age of AI. Market research becomes extraordinarily precise under AI. Purchase histories, search behavior, reviews, price elasticity. The quality of the analysis rises sharply. But precise analysis can only handle data that already exists. Data on a market that does not exist does not exist. If every company analyzes the same data with AI, every company arrives at the same conclusion. Beyond that conclusion there is no new market. The third conventional answer: “Innovation is a special project run by a dedicated unit” The third answer concerns organizational design. Create a newbusiness division. Build a separate outpost. Set up an internal venture scheme. Detach it from the logic of the existing business and protect it under a different set of measures. As early protection this design is correct. A business that has just been born will always fail the criteria of the established business. The design has a structural side effect. Innovation gets institutionalized as an exception. What is treated as an exception does not change decisions in the main body. Whatever happens at the outpost, the head office’s budget allocation criteria, appraisal scheme, and meeting agendas stay as they were. A few years later the outpost is cut back. The stated reason is usually that the core business is under pressure. The Enterprise Redefinition Maturity Model (ERMM) calls this Level 3, the Transformation Enterprise. Enterprise-wide initiatives do emerge. But transformation still occurs periodically, and organizations continue viewing redesign as a project rather than a permanent organizational capability. What the three conventional answers share is that they treat innovation as a special event. What the Age of AI puts in question is that assumption itself.

3 Redefinition — innovation is creating value and

bringing a market into existence Future Value Theory defines innovation as follows. Innovation is the work of redefining value and bringing into existence a market that does not yet exist. It is not releasing a new product. It is not advancing a technology. Those can be means. They are not the definition. Value is not discovered; it is created Begin with the most important shift. Most companies treat value as an object of discovery. An unmet need is buried somewhere, and whoever finds it first wins. Under that worldview, innovation is a search activity. Future Value Theory takes a different position. Value begins to exist at the moment somebody creates it. The smartphone was not an improved telephone. It rewrote the definition of “telephone” into a calculating machine and a service counter always at hand. After the rewrite, the markets for communications, payments, maps, photography, music, and advertising were rearranged. The electric vehicle is the same. It was not a replacement of the engine. It moved the definition of the automobile from a machine to a body of continuously updated software. The moment the property of being updated was added, the meaning of ownership, the mechanics of selling, and the point at which revenue arises all changed. Generative AI followed the same structure. It is not an improved search engine. It redefined information technology from a tool for looking things up into a tool for making things. The three share one feature: no prior market research produced these conclusions. Demand came afterward. Demand arose because value had been defined. This is the redefinition of Business, the second dimension of Enterprise Redefinition. Not “what do we sell” but “what value do we deliver.” Answering that question again is where innovation starts. Placing Christensen’s theory accurately The theory of disruptive innovation deserves an accurate statement here, because the concept is widely misused. Two kinds of innovation were distinguished (Christensen, 1997). Sustaining innovation improves a product along the performance axis that existing mainstream customers value. Strong companies are good at this. In sustaining competition, incumbents almost always win. Disruptive innovation is different. At the outset it underperforms existing products on the dimensions mainstream customers demand. It is superior on another axis. Cheaper, smaller, simpler, easier. So it establishes a foothold not in the mainstream market but among low-end customers or among non-consumers who could not buy before. There it improves, eventually reaches the performance level mainstream customers require, and takes the market. The core of the innovator’s dilemma sits here. Incumbents pass on disruption not because management is foolish, but because management is rational. They listen to their best customers and allocate resources to their most profitable businesses. That correct managerial behavior is what rejects investment in a disruptive technology. The remedy proposed was to separate the disruptive business into a distinct organization. So what does AI change, and what does it not? What changes is the speed of recognition and the frequency of disruption. Signs of substitution at the low end, the crossing point of performance curves, the size of the non-consuming population — AI makes all of these visible far earlier. At the same time, because development and validation get cheaper, the number of parties attempting disruption rises. The cycle of disruption shortens. What does not change is the dilemma itself. The dilemma is not caused by a shortage of information. It sits in the structure of resource allocation. The fact that the existing business earns more does not disappear under AI. Neither does the fact that the appraisal system is built on the metrics of the existing business. AI converts “we could not see it” into “we can see it and cannot move.” That is progress. It is not a solution. A problem of recognition can be solved by technology. A problem of will cannot. One further distinction. The theory of disruptive innovation assumes the existence of a performance axis. The creation of value is the work of establishing that axis in the first place. The former is largely a theory of defense and entry. The latter is a theory of creation. What becomes scarce in the Age of AI is the latter. Seeing societal challenges as Future Resources So where is a market that does not yet exist to be found? Future Value Theory’s answer is direct. Societal challenges. First Principle 7 states it. Social Challenges Are Future Opportunities. Social challenges are future opportunities — the origins of future markets, industries, and capital. The canon calls these Future Resources. They are handled as resources, not as problems. The renaming is not wordplay. It physically changes the entrance to innovation. A company that starts from “what will sell” looks at the data of existing markets. Every company holds the same data. Put AI on it, and the conclusions converge further. A company that starts from “which challenge do we take on” looks at places where there is no market yet. Aging, labor shortage, energy, food, gaps in education, keeping regions alive, access to care. There is no customer list there, and no demand forecast. But there are three properties. First, challenges do not run out. Second, there are few competitors, because nobody recognizes the space as a market. Third, the time horizon is long — which is why a company driven by short-term indicators cannot enter. Societal challenges as Future Resources are also the starting point of the Future Value Cycle. Societal Challenges → Purpose → Future Value → Enterprise Value → Capital

→ New Challenges → Societal Progress → Greater Future Value

In this cycle, solving is not the endpoint. The capital earned by solving becomes the funding for the next challenge. Value in this cycle is regenerative rather than linear. A company whose innovation ends as a one-off is not turning this cycle.

4 Structure — at which stage new value arises

Bring the redefinition down to a working skeleton.

4.1 Creation is the fourth stage of the chain

The Future Value Chain gives the causal order of value creation. Purpose → Learning → Redefinition → Creation → Enterprise Value Innovation occurs at the fourth stage, Creation. The decisive point is that Creation does not start on its own. Without Purpose, what should be created is undecided. Without Learning, there is no material to create from. Without Redefinition, the company tries to handle new value in its old shape. And enterprise value appears last, as the outcome of the chain. Most corporate new-business efforts begin at Creation. Search for a theme, list proposals, pick one, build it. The three upstream stages are missing. So the proposals that emerge look like something seen elsewhere, the criteria for choosing never settle, and the effort runs out of strength before completion. Before the idea there is a question. Before the question there is a purpose. In an era when AI carries generation, the order matters more. When generation is free, only the quality of the upstream makes a difference.

4.2 Why new businesses keep failing

Using the chain, the structure of failure breaks into five parts. First, a missing origin. Purpose is skipped and the work begins with a theme search. Because what it is all for is never settled, every judgment along the way becomes a vote. Second, a contaminated axis of evaluation. The new business is measured by the criteria of the existing business. Profit in year one, an accurate demand forecast, a certain payback period. A newborn business satisfies none of the three. Demanding new businesses without changing the criteria is a contradictory instruction. Third, a shortage of time. The Value Equation shows this directly. Value = Purpose × Trust × Capability × Time This is multiplication. It is not addition. If Time is zero, value is zero however high the other three. A business with a clear purpose, real capability, and trust produces no value if it is shut down after two years. Equally, if Trust is zero, however superior the technology, society does not accept it. Because the relationship is multiplicative, value without purpose has no direction, without trust cannot spread through society, without capability cannot be realized, and without time cannot endure. Fourth, solo development. The company tries to complete everything internally, without an ecosystem. The FVCC Formula holds Ecosystem as an independent term. FVCC = Purpose × Learning × Redefinition × AI Integration × Ecosystem × Capital Allocation × Trust This is multiplication too. A company with Ecosystem at zero has zero capability to create Future Value however high the other six. The relationship is multiplicative: weakness in any single capability weakens the whole, so purpose, AI, and capital are each individually insufficient, and Future Value emerges only when all seven reinforce one another. A new market does not stand up without customers, regulators, suppliers, universities, and local government. Alone, a company can build a product. It cannot build a market. Fifth, the disappearance of learning. On exit, the team is dissolved and no record is kept. What was learned from the failure does not remain in the organization, so the same failure repeats three years later. This is why Learning sits at the second stage. Failure itself is not the problem. A failure that is not converted into learning is the problem.

4.3 Designing choice as a system

In a world with a surplus of ideas, choosing becomes the largest act of management. First Principle 3 supplies the criterion. Capital Exists to Create Possibility. Capital exists to create possibility, not merely to maximize return. On that criterion the investment question changes. It is not “how much will this proposal earn.” It is “if this proposal succeeds, what possibility comes into existence for society.” AI can compute the first. Only a human can settle the second. This is where Human-on-the-Loop Management becomes necessary. Management that reviews proposals one at a time breaks down the moment proposals multiply a hundredfold. What is needed is to design the rule that decides which proposals pass and which do not. Not the individual judgments, but the conditions under which judgments are generated. That is what standing above the loop means. Human beings sit above the system, responsible for designing the whole rather than controlling each action; the objective is better design rather than better control.

5 What it looks like in practice — what is happening on

the ground Lay the theory over specific industries and organizations. What remained in the industries where search was solved In drug discovery, the search for candidate molecules has been greatly accelerated by AI. Screening that once took years finishes in far less time. Has the number of new drugs risen by the same multiple? It has not. The bottleneck moved. What remains is the design and conduct of clinical trials, the dialogue with regulators, the start-up of manufacturing, and implementation in clinical practice. Every one of these requires time, trust, and social agreement. The same thing is happening in materials development. AI covers the search space of compositions. What remains is scaling to volume production, and the work of building a new application market for the material. The structure is common. AI solves search. It does not solve implementation. And value does not exist until it is implemented. What a company that enters from a societal challenge looks like Put the entrance at a societal challenge, and the shape of the business changes. Consider a logistics operator that treats labor shortage as a Future Resource. The question is not “how do we make delivery more efficient.” It is “how do we build a system by which society functions with fewer people.” That brings joint delivery, demand leveling, and regional partnership into range alongside automation. A company that treats aging as a Future Resource handles the burden of care not as a product specification but as a design problem for society. It tries to solve at a unit that includes medical institutions, local government, families, and the insurance system. An ecosystem becomes unavoidable here. Neither conception comes out of market research. The market does not exist yet. The challenge, however, is in front of us now. The meeting room where ideas multiplied a hundredfold One more scene, currently playing out in many companies, deserves recording. A new-business unit deploys AI and becomes able to generate proposals in bulk. A hundred proposals arrive at the meeting. What happens next? Nothing gets decided. The reason is simple. There is no criterion for choosing. In an organization without a criterion, more options make comparison harder and push the decision back. In the end, the proposal closest to the existing business is selected, because it attracts the least opposition. AI amplifies the organization’s weakness. In an organization with a clear purpose, AI is a device that widens the options. In an organization with a vague purpose, AI is a device that worsens the inability to decide. The same tool works in opposite directions. When failure becomes a resource The failure rate of new businesses is high. That is not a defect of the system; it is the nature of work that handles the unknown. So the question is not how to reduce failure. It is how to convert failure into a resource. In the Future Value Chain, Learning is placed immediately after Purpose. The material for learning comes more from failure than from success. Only when a hypothesis is wrong can the error in the assumption be identified. In most companies, though, failure does not become a resource. There are three reasons. First, failure is tied to the individual’s appraisal. The moment it is tied, the people involved report the failure as smaller than it was. Second, the hypothesis to be tested was never written down in advance. If it was not written, what was refuted cannot be known. Third, no record of the exit is kept. Put the other way, design those three and failure becomes an asset of the organization. Write the hypothesis first. Tie the result to the hypothesis, not to the person. On exit, record separately what was validated and what was left untested. That alone sharply reduces the repetition of the same failure. AI does make the testing of a hypothesis faster. It cannot decide which hypothesis to form. Here too, what gets faster is validation, not choice. What an organization that made creation ordinary looks like Organizations that have moved innovation from a special project to the normal condition share four features. First, the question is open to the whole company. “Which of our assumptions is becoming obsolete?” is asked not only by the newbusiness unit but by the front line of the existing business. Second, capital allocation is continuous. Instead of one annual budget meeting, the allocation is rebuilt each quarter. Increases, reductions, and stoppages are handled in the same meeting. Third, exit is designed as learning. On exit, what was validated and what remained untested are recorded. The people are not dispersed; they are moved to the next attempt. Fourth, the existing business also carries a duty to create. The division of labor in which the existing business defends and the new business attacks has been abandoned. In the language of the ERMM this is the move from Level 3 to Level 4. At Level 4, the Continuous Redefinition Enterprise, enterprise redesign becomes embedded within normal management processes. And organizations increasingly redesign themselves before external disruption requires it. Three cautions must accompany any use of the model. Progression is not linear: organizations frequently display characteristics from multiple levels simultaneously, and a firm may possess Level

4 AI capability while remaining Level 2 in leadership. Maturity is

assessed across all five dimensions in balance: exceptional technological capability with weak leadership redesign cannot achieve higher maturity, and strong purpose without adaptive organizational systems remains insufficient. And reaching Level 5 as rapidly as possible is not the objective, because different industries may require different levels of organizational adaptability. What matters is not the height of the level but the coherence of the five dimensions.

6 Questions for the executive

The argument, in one line. Innovation in the Age of AI is not the production of ideas. It is the work of redefining value, taking on societal challenges as Future Resources, and continuously bringing into existence markets that do not yet exist. AI carries the ideas. Human beings carry the choosing, the investing, the building, and the will to carry it into society. As First Principle 4 states, AI optimizes; humans define value, purpose, and direction. Three questions to close. Each can be answered at your next executive meeting. Question 1 — Does your new business begin from a purpose, or from a theme search? If it begins from a theme search, the business began at Creation. The three upstream stages are missing. A business run without them will always come back to the question of why it is being done. If there is no answer then, the business stops quietly. Question 2 — When ideas multiply a hundredfold, on what criterion will you select three? Can you write that criterion down in words? If you cannot, deploying AI will slow your decisions. For an organization without a criterion, an increase in options is not an asset. It is a liability. Question 3 — Can you name the societal challenge your company has taken on? If you cannot, your innovation happens only inside existing markets. Inside existing markets, AI is converging every company on the same answer. Effort in a place where no difference can appear produces no difference. None of the three questions asks about technology. All three ask where value is raised from. First Principle 6 states that Enterprise Exists to Redefine Itself. Continuous self-redefinition is its essence. Innovation is nothing other than that redefinition appearing on the outside. So innovation is not the work of a new-business unit. It is the mode of existence of the enterprise itself. In an era when AI supplies ideas without limit, what is scarce is not conception. It is deciding which future to bring into existence, betting capital and time on that decision, and carrying it all the way into society. AI does not have that will. The executive does.

In brief

Innovation in the Age of AI is the work of redefining value and bringing into existence a market that does not yet exist.
Value is not an object of discovery. It begins to exist the moment somebody creates it, and demand comes afterward.
AI multiplies ideas without limit. What becomes scarce is the criterion for choosing and the will to bet.
Societal challenges are Future Resources. A new market can only come from a place where there is no market.

Key concepts

Future Resource / Future Value Cycle / Enterprise Redefinition / Future Value

The chain of ideas

Future Resource → Purpose → Redefinition → Creation → Future Value

Related first principles

Principle 3 — Capital Exists to Create Possibility. Principle 6 — Enterprise Exists to Redefine Itself. Principle 7 — Social Challenges Are Future Opportunities.

Related chapters

Vol. I, Ch. 006 “What Is Competitive Advantage in the Age of AI?” — why democratized capability stops being a difference
Vol. II, Ch. 013 “What Is Corporate Culture in the Age of AI?” — the conditions under which creation stops being an exception
Vol. V, Ch. 046 “What Does It Mean to Redefine a Business Model?” — how redefinition of the Business dimension proceeds
Vol. VII, Ch. 069 “Does Innovation Become Enterprise Value?” — how creation is measured as value

Papers and companion volumes

Kadowaki, N. (2026a). Future Value Theory: A Management Framework for Enterprise, Capital, and Society in the Age of AI. VURA Working Paper Series. SSRN: https://ssrn.com/abstract=7120980 / Zenodo: https://doi.org/10.5281/zenodo. 21255662
Kadowaki, N. (2026b). Enterprise Redefinition: Toward an Enterprise Evolution Theory for the Age of AI. VURA Working Paper Series. (Published on Zenodo; under review at SSRN)
100 Questions on Management in the Age of AI, #051 “Why Do New Businesses Keep Failing?” / #053 “Who Creates Innovation in the Age of AI?” / #055 “Is Failure Necessary in the Age of AI?”

Chapter 014 What Is Innovation in the Age of AI?