Every growth framework is a search method. Each one looks in a different place.
At Exaptation AI, we think of growth as a search problem. Every founder, executive, and investor is searching for the next source of growth, and every growth framework is a way to search. The quickest way to tell the frameworks apart is to ask where the search starts.
Here is a recent example. In a conversation with Replit's Amjad Masad this past May, Stripe co-founder and CEO Patrick Collison was asked how he would try to get into entrepreneurship today, and what some practical ideas would be. His answer had three parts. Start trying things right away. Go after industries still running on clunky 1990s software. And look at what is popular with young people but still low status.
That is useful advice. It is also three different searches. The first is about how to search. The second starts from what incumbents do poorly. The third starts from how people behave. None of them starts from what a company already owns, and that is the search we care most about.
Four frames
Academics and most practitioners already know three of the big frames for this, even if they don't always tell them apart. With Demand Pull, the market pulls new products into existence. With Supply Push, a firm pushes what it already makes or does well toward the market. With Technology Push, a new capability pushes its way into products and markets.
Supply Push and Technology Push are often treated as the same thing, and parts of the literature use the two terms interchangeably. We separate them because the difference matters in practice, especially between incumbents and new entrants. Established firms tend to push what they already have. New entrants tend to push something new.
There is a fourth place to start. We call it Resource Expansion. It borrows from all three frames and adds exaptation logic to strategic management. Exaptation is a term from evolutionary biology for a trait that takes on a function it was not originally shaped for. In business, it means an asset or resource taking on a function it was never designed to perform. This fourth frame is not used much in the existing literature. That is the point. It is where Exaptation AI does its search.
Figure 1: Four places to start a growth search. How to read it: each column is one frame, read top to bottom: where the search starts, the frameworks that live there, and a one-line example. Resource Expansion is marked in gold because it is the frame Exaptation AI works in. The band below gives its definition.
Exaptation AI · Growth is a search problem · v0.2
Four places to start a growth search
Every growth framework is a search. The quickest way to tell them apart is to ask where the search starts.
Demand Pull
Also called Market Pull
Starts fromWhat customers or markets want
Disruptive Innovation
Christensen, 1997
Steel minimills entered with cheap rebar, then moved upmarket
Blue Ocean Strategy
Kim & Mauborgne, 2005
Cirque du Soleil: a circus without animals, priced like theater
Jobs to Be Done
Ulwick; Christensen et al., 2016
Commuters “hired” a milkshake to get through a long drive
Supply Push
Starts fromWhat you already make or do well
Resource-Based View
Barney, 1991
Honda carried its engine know-how into motorcycles, cars, lawn mowers and generators
Effectuation
Sarasvathy, 2001
A cook with a few recipes could end up with a caterer or a food brand
Technology Push
Starts fromA capability that is new to you
General Purpose Technology
Bresnahan & Trajtenberg, 1995
Factories gained from the electric motor only after redesigning around it
Today’s version
AI, and most of the growth advice that comes with it
Resource Expansion
Also called Asset Expansion
Starts fromWhat else the things you own or can access can do
Exaptation Growth Strategy
Shah, 2024
Uber’s network, built to move people, started delivering food
Getting more value from what you already own or can access, by expanding what it does or finding functions it was never designed for.
Each framework is placed by where its search starts, not by its full content. Supply Push keeps a resource in its designed function. Resource Expansion looks for functions it was never designed to perform.
1. Demand Pull (also called Market Pull)
The search starts from what customers or markets want. Disruptive Innovation (Christensen, 1997) looks at the edge of the market, at overserved low-end customers and nonconsumers. Steel minimills entered with cheap rebar and then moved upmarket. Blue Ocean Strategy (Kim and Mauborgne, 2005) looks across industry boundaries for uncontested space, the way Cirque du Soleil built a circus without animals and priced it like theater. Jobs to Be Done looks at the customer's circumstance and the job they are struggling to get done. In Christensen's well-known example, commuters “hired” a milkshake to get through a long drive.
Here the job belongs to the customer. That is why we use a different word, function, for what an asset does.
2. Supply Push
The search starts from what a firm already makes, holds, or does well, pushed toward the market in its designed function. The resource-based view (Barney, 1991) explains why this works: resources that are valuable, rare, and hard to copy are a lasting source of advantage, so firms push them into more products and markets. Honda carried its know-how in small engines into motorcycles, cars, lawn mowers, and generators, the example Prahalad and Hamel made famous in their 1990 work on core competencies.
Effectuation (Sarasvathy, 2001) starts in a similar place, with the means in hand: who you are, what you know, and whom you know. A cook with a few recipes could end up with a caterer or a food brand.
Supply Push keeps the resource in its designed function. More of it, done better, for more customers, or in more markets.
3. Technology Push
The search starts from a capability that is new to you. The best-known theory here is General Purpose Technology, from Tim Bresnahan and Manuel Trajtenberg (1995). A few technologies, like the steam engine, electricity, and computing, spread across many sectors and raise productivity, but often only after a lot of complementary investment. Factories gained from the electric motor only after they were redesigned around it.
AI is today's version, and many of the growth ideas we track fit within this construct. This is also where most new entrants start. There will be differences between what the theory says and how it is applied with new tech like AI, and the tracker will show some of them.
4. Resource Expansion (also called Asset Expansion)
The search starts from what you already own or can access, and asks what else it can do. We define it as getting more value from what you already own or can access, by expanding what it does or finding functions it was never designed for.
Exaptation Growth Strategy, the framework behind Exaptation AI, searches the assets and resources a firm owns or can access for functions they were never designed to perform. Uber's network was built to move people. Then it started delivering food, and in 2025 Uber reported $17.2 billion in Delivery revenue.
The difference from Supply Push is the function. Supply Push keeps a resource in its designed function. Resource Expansion looks for an emergent function, one the resource was never designed to perform.
This is not about doing more with less. It is about doing more with more: the potential already sitting inside what a company has built.
Why the fourth frame gets the least attention
We think Resource Expansion is the least searched and least mentioned of the four. The first release of our tracker supports that so far. None of the 11 classified signals from CEOs and investors points to Resource Expansion, and one of the 13 requests in Y Combinator's latest full list does.
A few reasons seem likely. Much of the public advice about growth is aimed at founders, who start with few assets to look at. A new technology arrives with a launch event and a press cycle. An old asset's undiscovered function arrives with nothing. And we tend to know the things we own by the function they already perform, which makes their other functions hard to see.
There is evidence that the raw material is there. In a large survey of European inventors from the mid-2000s (PatVal-EU), about 36% of patents were not used at all, and large firms left more unused than small ones.
Our working view is that this frame is underutilized, less competitive, and a hidden advantage for people who want to try it first. The other claims will be proven by data as we collect it.
A hypothesis about the frame underneath
General Purpose Technology may operate below the surface of the other three frames, helping generate growth from the demand side, the supply side, or the resource side. AI changes what customers can be offered. It also changes what existing assets can do. a16z's Jonathan Lai expects game worlds built for play to double as training grounds for robots, which would be an emergent function made possible by a new technology. This may be controversial academically, but it may be more apt in practitioner circles. We will watch for it in the data.
Introducing Growth Search Signals
Today we are launching Growth Search Signals, a tracker of where influential people and capital allocators say to search for growth. We review their public conversations, speeches, and writings, check each one against the original source, and use their own words to sort each signal by where it tells you to start searching.
Release 1 holds 25 signals: the CEOs of Amazon, Microsoft, Nvidia, OpenAI, and Stripe, partners at a16z, and Y Combinator's Fall 2026 requests for startups, including one from the US Army. Each entry shows who said it, the growth thesis in one sentence, a dated reason why now, the frame, the closest theory, and an ExAI note. The note ends with where we would start, named from our own library of causes and exaptation experiments. The information is easy to find. That last line is not.
Today's count from CEOs and investors is 4 Demand Pull, 2 Supply Push, 5 Technology Push, and 0 Resource Expansion, with one signal left unclassified. It shows work in progress instead of trying to force-fit everything into neat categories. We count Y Combinator's requests as their own series, so one accelerator doesn't drive the headline.
Two things are worth saying up front. We are not a neutral party. We build tools for the fourth frame, so everything is in the open for readers to disagree with and debate. And we have made a public bet: we expect Resource Expansion to stay the least searched frame, and we will report where it stands at 50 signals, whatever it shows.
The tracker is provisional and built in public. We will update it every two weeks or so. Next, we will add a companion tracker of the bets startups actually make, both as new entrants and when they pivot. One shows where capital allocators say to search. The other shows where startups are actually searching for growth and how.
Every search pays
Figure 2: Every search pays. How to read it: the three blue cards expand the function an asset was designed for: performed better, for new customer needs, or in new markets. The gold card is an emergent function, one the asset was never designed to perform. One Uber example illustrates each kind.
Exaptation AI · Growth is a search problem · v0.2
Every search pays
Searching what you already own returns four kinds of growth. Three expand the function an asset was designed for. The fourth is a function it was never designed to perform.
Expanding the designed function
1 · Do it better
The same function, performed better
Uber: faster pickups and better matching for riders
2 · New use cases
The same function for new customer needs
Uber: business accounts for company travel
3 · New markets
The same function in new markets
Uber: rides in new cities and countries
Finding an emergent function
4 · New function
A function it was never designed to perform
Uber: the network built to move people starts delivering food
The Uber lines illustrate each kind of return. Food delivery (Uber Eats) is EGS Case Study No. 3. When the first three are the goal, that is Supply Push. The fourth is rarer.
Growth is a search problem, and no single frame wins every search.
Demand Pull, Supply Push, and Technology Push will keep producing most of the ideas, and they should. Our contribution is one new and significant way to search for growth, aimed at the place that gets the least attention.
A search of what you own usually returns more than one kind of value. Along the way you find ways to improve the designed function, new use cases, and new markets. When those are the goal, that is Supply Push, and they are worth having. The rarer find is an emergent function.
Every search pays. You see what more your assets can do today, on the way to functions they were never designed for.
At Exaptation AI, we are building the search engine for every asset's undiscovered functions. The question we keep asking is simple: what else can the things we already own do?
See the trackers: Growth Signals