From Crowds to Algorithms: Why Your Company Should Bet on Itself

Public Forecasting: Markets Meet the Ballot

Election season 2026 has been less “who will win” and more “how much money is riding on who will win,” which, depending on who you ask, is either a fascinating new information source or a five-alarm fire for the integrity of the process. “Betting” volume on the midterms has exploded, with tens of millions of dollars sitting on contracts about which party controls Congress (AP News, 2026). Political scientists like the real-time read on public sentiment, as markets often incorporate new information faster than polls (Wolfers & Zitzewitz, 2004). State election officials like it considerably less, mostly because “people are financially incentivized to influence the outcome” is not usually a sentence you want anywhere near the word “election.”

Even Congress became nervous enough to ban its own members and staff from trading on these markets, which is either a sensible ethics rule or the most backhanded compliment prediction markets have ever received (AP News, 2026). “This tool is clearly good at revealing information. Insiders, please stop using it.”

Gaming and Gambling: The Regulatory Whack-a-Mole

Here’s where it gets genuinely funny, in a “several state attorneys general are extremely mad” kind of way. About 90% of one platform’s volume is sports bets (Pew Research Center, 2026). The platforms insist, with a straight face, that this is not gambling, it’s “price discovery.” Regulators in New York, Michigan, Arizona, Portugal, Hungary, and India have each responded with a version of “sure, and I have a bridge to sell you,” and started issuing cease-and-desist letters (Courthouse News Service, 2026). New York’s governor summed up the regulatory mood in one sentence: gamble with our laws, and you’re going to lose. Meanwhile, the federal government is suing its own states to stop them from enforcing gambling law against the platforms, which is a sentence that should not make sense and yet, 2026.

The moral for management types: if you build an internal version of this, please, for the love of your compliance department, use fake money. “Prediction market” and “gambling” are separated mainly by currency and vibes, and vibes don’t hold up in court.

Trading: The Robots Have Entered the Chat

While regulators argue about what to call all this, AI agents have quietly started trading prediction markets on their own-24/7, no lunch breaks, no vacation days, no existential crisis about whether the Fed will cut rates. This makes a strange kind of sense: these contracts are just natural-language questions (“Will X happen by Y date?”), and reading natural language is precisely the party trick large language models are best at (Halawi et al., 2024).

One study found prediction markets beat Wall Street analysts at forecasting earnings the day before results dropped, largely because markets correct analyst biases in real-time (Rabetti et al., 2026). Even JPMorgan is reportedly building its own internal version, notably, one that deliberately excludes sports and politics, because Jamie Dimon apparently does not need a market on whether the Fed chair keeps their job to know that’s a distraction (CoinDesk, 2026).

So Why Isn’t Your Company Doing This Already?

Here’s the part where the story gets a little embarrassing for corporate management as a discipline: we already knew this worked, twenty years ago, and then mostly forgot about it.

Hewlett-Packard ran internal markets in the late 1990s to forecast printer sales and beat its own official forecasts most of the time (Chen & Plott, 2002). Google ran internal markets on launch dates for years, finding that employee trading provided more accurate probability assessments than executive timelines (Cowgill & Zitzewitz, 2015). Ford, Microsoft, and Best Buy all poked at the idea. And then... it mostly stayed a fun HR anecdote instead of becoming a management discipline. Why? Nobody wanted to run it. Internal markets need active moderation, contract design, and someone willing to explain to Legal why employees are “trading” on whether the product ships on time (Cowgill & Zitzewitz, 2015). It’s a little insulting to management, honestly. A market that continuously and transparently prices your department’s chance of hitting a deadline is not always a popular addition to the org chart, even if, especially if it’s more accurate than the status update in the Monday meeting.

Prediction markets have a reputation problem. When the loudest headlines about the technology are “state sues company over illegal sports betting” (Courthouse News Service, 2026), it’s hard to pitch your CFO on the internal, virtual-currency, entirely-legal version without a lot of caveats up front.

Enter AI, Stage Left

This is the actual argument of the (much longer, much more citation-heavy) paper this post draws from stated as follows: AI quietly fixes all three problems.

Automation: An AI agent can run the market-reading internal updates, news, and data, and estimating probabilities—without a human having to babysit it or design every contract by hand. Recent research confirms that retrieval-augmented AI systems can now approach human-level forecasting accuracy on complex questions (Halawi et al., 2024).

Reframing: Framing it as “decision support,” not “a machine that publicly grades your forecasting,” makes it a tool management uses rather than a tool that replaces management. Much easier to sell in a budget meeting.

Legality: Keep it on virtual currency, keep it internal, and you dodge the entire gambling-regulation circus currently playing out in eleven state courts simultaneously.

In other words, the technology that’s currently making regulators write strongly worded press releases about sports betting is the same technology that could quietly make your quarterly forecasts a lot less wrong, as long as you keep it far, far away from anything resembling a sportsbook. Somewhere between “Congress banning its own staff from betting on Congress” and “an AI agent trading 24/7 so it doesn't have to sleep,” there’s a genuinely useful management lesson buried in all this 2026 chaos: markets are honest in a way meetings are not, and AI has just made that honesty a lot cheaper to install.

Whether your management team is ready to hear that from a betting slip is, admittedly, a separate question.

Practitioner Deep Dives, A Manifesto for Radical Honesty

True leadership is not found in the comfortable consensus of a boardroom, where optimism is mandatory and doubt is punished. It is found in the gritty, uncomfortable pursuit of what is actually true, regardless of whose ego gets bruised in the process. This is the fundamental ethos that underpins The Leadership Pantheon: Practitioner Deep Dives. This series rejects the sanitized narratives of corporate success, where every project is “on track” until the day it collapses, and instead demands an unapologetic autopsy of our decision-making delusions. We aim to bridge the chasm between the polished executive summary and the messy, chaotic reality of the operational floor. By exploring tools like prediction markets, we are not just suggesting a new piece of software; we are challenging the systemic gaslighting that allows organizations to lie to themselves until they run out of cash. We are here to drag the structural absurdities of modern management into the light, exposing the comfortable fictions leaders tell themselves to sleep at night.

Ultimately, this series compels us to finally confront the pivotal Leadership Pantheon Reflection Question: “Are you shielding your strategy behind the polite fiction of executive intuition, or do you have the leadership fortitude to let the cold, hard odds dismantle your ego for the sake of the mission?”