AI Disruption Has Plenty of Opinions. Now It Has a Prediction Market.

AI Disruption Has Plenty of Opinions. Now It Has a Prediction Market.

Everyone’s heard the Sherlock Holmes story about the dog that didn’t bark. The one where a champion racehorse disappears and its trainer turns up dead. Everyone’s poking around the crime scene trying to find a clue to solve the mystery. Holmes eventually does by focusing on what didn’t happen. As in the stable dog that never barked. It was the clue that cracked the case.

One might say that AI has a lot of dogs that aren’t barking, too.

There’s the doctor’s appointment that never gets booked because a chatbot answers the question at 11 p.m. The support ticket that never gets opened. The junior analyst who never gets hired because three people can now do what took five. The software subscription that was never going to renew because the work moved somewhere else.

None of those decisions leaves an immediate paper trail or a receipt, so to speak. But every one of them says something important about what AI is doing to the economy.

That’s what makes measuring a little AI tricky. Everyone can see the money pouring in. Chips. Data centers. Models. Talent. What’s much harder to see is what comes out the other side. What changed, what disappeared, what got faster and what became unnecessary.

Starting in September, people will be able to trade on what’s otherwise been so very hard to put a price tag on.

Today, PYMNTS and Crypto.com are announcing AI Prediction Market Contracts. More than 20 contracts will be listed at launch, with roughly 25 more added every quarter. They will trade on OG Prediction Markets, the CFTC-regulated exchange and clearinghouse affiliated with Crypto.com. Distribution will come through Crypto.com’s introducing broker and other partners. This is a two-year exclusive partnership.

Read More: Crypto.com and OG.com Turn AI Disruption Into AI Predictions Market Contracts

Every contract will settle against PYMNTS Intelligence data, built over nearly 30 months of continuous measurement across consumers, enterprises and public companies, with more added every month. The idea is simple. Put a price on the part of the AI economy everyone is trying to see but no one really has complete clarity around just yet. The prediction market will tell investors and business leaders what our research will find before it sees the light of day.

In other words, it will identify those dogs that don’t bark but offer incredible clues to how the AI economy will evolve.

The AI Economy Is Full of Missing Receipts

Most technology waves are pretty easy to spot. Someone buys a terminal, a server, a software license or an iPhone. Count the purchases, the receipts, and you have a decent idea of where things stand.

AI and AI disruption doesn’t always work that way. Much of its impact shows up as something that no longer happens, which means it disappears from the data most people use to track the economy. You can’t count a support call that was never made. You have to ask what the customer did instead.

That’s why PYMNTS Intelligence asks 4,000 U.S. consumers every month what they actually did. The sample is balanced across 180 census quotas, so it is statistically representative of U.S. consumers. We’re not asking whether people are excited, worried or impressed by AI. There are more than enough polls for that. We’re measuring behavior. What did AI replace? What did it speed up? What did it persuade someone to buy, skip or do differently? How persistent and permanent is that shift?

If AI changes the decision but no conventional transaction records the change, asking is the only way to find the dog that hasn’t barked.

The Average Is Where the Real Signal Goes to Die

In the 1950s the U.S. Airforce wanted to standardize the design of airplane cockpits. To do that, it assembled a number of physical dimensions from more than 4,000 pilots to support the design for the “average pilot.”  A subsequent study later concluded that zero pilots fit the average across the dimensions studied in the survey. The Airforce scuttled its plans and reverted to a more flexible cockpit design.

Our years of looking at data on AI use and adoption concludes that there isn’t one AI “average” adoption curve. (Even though we shared one a year ago.) There are dozens, and they’re moving at many different speeds.

Read More: Gen AI: The Technology That Broke the Adoption Curve

Consumers are ahead of businesses in some categories and years behind them in others. Two companies in the same industry can be on opposite sides of the curve. Inside the same company, one department may be rebuilding work around AI while the group down the hall is still figuring out where it fits.

Blend all of that into one headline adoption rate and the useful information gets watered down. The average tells you there’s an AI wave. Thank you, but we knew that. After all, we do read PYMNTS.

The real signal is in the gaps, those dogs that haven’t barked yet. Which categories broke away from the pack this quarter? Where is consumer behavior running ahead of what companies have built, creating a demand gap someone will fill? Where is enterprise deployment ahead of business use, leaving a lot of spending in search of a use case? And where does a curve that was climbing fast look like it has lost steam, or worse yet, run out of it?

Read More: The Battle for AI Isn’t About Models. It’s About Habits

Those are the nuances, the edges, the precision that matters to business executives and investors.

For that reason, PYMNTS Intelligence has created a very sophisticated methodology to measure AI disruption to support the AI Prediction Market Contracts that will trade on Crypto.com.

Consumers are measured monthly. Enterprises and the executives who control AI budgets are measured quarterly. Corporate disclosures are tracked quarterly through structured analysis of SEC filings and earnings materials. The interesting part will be where one of those groups moves faster than the others. And those are the contracts we will field, and where the real signals will be found.

Behavior Moves First. Earnings Move Last.

Over the last 30 months, we’ve also observed that AI’s economic impact tends to travel in a rolling wave.  Consumers change what they do first because they don’t need a budget, a procurement process or permission from a steering committee. With their free apps, to start, they can experiment at a velocity that businesses can only dream about. The tools are sitting on their phones. Their instructions are coming a mile a minute at their prompts.

Read More: Why 30 Million US Consumers No Longer Search

Companies deploy next. They disclose later, usually in language reviewed and edited until it says just enough to hint there’s something being done. The impact reaches earnings last. By then, it’s often already baked in the stock price, and everyone is reading history. That lag can run for several quarters. It’s also where we believe the opportunity for clarity lives.

Read More: The AI You’re Given vs. the AI You Chose

For example, semiconductor demand is a useful measure of what the AI buildout costs. It can’t tell you whether the buildout is changing behavior or making anyone money.

You want to see pharmacy visits being displaced before a pharmacy chain lowers its guidance. You want to know that enterprise agents have plans to replace their software with AI models before a software company’s renewal rate delivers the news. Those signals exist months earlier if someone is measuring behavior often enough to catch them.

And we are.

A Number Tells You Where Things Are. A Price Tells You Where They May Go.

Research gives you a reading at a point in time. We do and publish a lot of that. A market prices what people think will happen next.

For example, if a contract tied to a specific AI adoption threshold trades at 30 cents and then moves to 70 cents over six weeks, something’s changed. The market now assigns a much higher probability to that threshold being crossed. It also shows how quickly conviction moved and how much money stands behind the new view.

That’s a forecast that updates whenever new information hits. A static number can’t do that. Through trading on contracts, investors will provide a real-time read on what our surveys will find. That’s what prediction markets do. Just like you won’t have to wait until the votes have been counted to find out who’s likely to win an election, you won’t need PYMNTS to reveal its research to have a market-based estimate of where it is going to land.

These binary contracts force everyone to get specific. “AI is transforming healthcare” is just more blah blah blah. A contract with a defined population, threshold and settlement date turns the talking point into a claim about a particular outcome at a particular time.

The prediction market adds another layer. Every participant brings a view from some corner of the economy. A CFO knows what her own deployment looks like. An operator sees renewal behavior inside a category. An analyst follows a supply chain few others watch. Those partial views get pushed into one live price.

We’ve seen AI disruption and its impact on payments, commerce and the economy at large. What was missing was a market that could price it.

Well, until now.

A New Kind of AI Information Market

The AI Prediction Market Contracts aren’t another AI report, dashboard or index. It’s an information market that converts independently measured AI adoption and impact into tradable binary contracts.

The contracts are regulated event contracts listed on the Crypto.com exchange, which is registered with the CFTC as both a designated contract market and derivatives clearing organization. They settle against published, source-traceable and reproducible readings from PYMNTS Intelligence.

The better comparison is the interest rate futures market. It exists because businesses and investors needed a continuous price on something everyone had a view about, but no one could know with certainty.

The AI economy now has a version of that problem.

The settlement data also can’t be moved by one market participant. Readings come from a nationally representative survey of 4,000 U.S. consumers, a 500-firm enterprise study and verified corporate disclosures. No trader can change what 4,000 Americans or 500 of the largest companies in the U.S. did last month.

More than 20 contracts will be available when trading opens in September, with roughly 25 more added each quarter. The first group goes where AI’s impact is both hard to see and valuable to price. Consumer behavior across health, learning, work and commerce; enterprise deployment and workforce change; and shifts inside financial services, retail and healthcare.

More sectors will follow as the partnership expands into new technologies, industries and applications.

Complete methodology, contract specifications and the settlement calendar will be published before trading opens.

Every Transformation Eventually Gets a Price

AI may be the biggest capital allocation decision of our lifetimes, and much of it is still being made on conviction. Right now, AI’s impact is unusually hard to see, but at the same time, something everyone knows is real. It hides in things that don’t happen, gets flattened by averages and reaches earnings long after it reaches the real world.

Read More: How Leading Enterprises Really Measure Gen AI ROI

In September, those signals get a continuously updating market price.

Everybody has an opinion about AI. We’re about to find out what those opinions are worth.

And how loudly those dogs are finally barking.

I hope you are as excited about this as our collective teams at Crypto.com and PYMNTS are. It’s been an extraordinary journey, and yet we’ve only just begun.

To get more of the details, and to sign up for alerts when contracts go live, please visit the AI Prediction Market.

The post AI Disruption Has Plenty of Opinions. Now It Has a Prediction Market. appeared first on PYMNTS.com.