Artificial intelligence (AI) is changing how crypto and traditional markets get traded, yet four leading analysts agree it rewards skill rather than replacing it. The edge in AI in crypto trading still comes from clean data and human judgment.
Charles Edwards of Capriole Investments and Julio Moreno of CryptoQuant call AI an accelerant for serious research. Benjamin Cowen and Michael van de Poppe, speaking on a separate panel, reach the same conclusion from the trading desk.
Four Analysts, One Conclusion
On-chain analytics and AI tools have moved from niche to mainstream across crypto research. Two BeInCrypto panels gathered four analysts who use them every day.
Edwards founded Capriole Investments, a quantitative Bitcoin (BTC) hedge fund. Moreno serves as Head of Research at CryptoQuant. Cowen and van de Poppe are widely followed, independent market analysts.
Speaking at the Market Intelligence Council, Edwards said AI shifts the opportunity toward those who do the work.
“I think AI as well is making that… playing field more opportunistic for certain people.”
On a separate panel, van de Poppe set the limit plainly.
“It’s not going to make you a great trader if you weren’t a good trader in the first place.”
Where AI Already Helps
The clearest gains show up in routine research. AI now compresses tasks that once took hours.
Edwards pointed to faster analysis as the main benefit.
“The tool sets to do that are much more powerful and… it can be done more quickly today with AI.”
Van de Poppe showed how accessible this has become. He built a sample crypto portfolio using a chatbot and free data feeds. Tools like AI agents now pull live market data on demand.
“You can build a portfolio and a dashboard of cryptocurrencies within five minutes with just free APIs.”
Why AI Still Needs a Human
Speed does not equal skill. Van de Poppe noted that his AI portfolio missed important context.
“It didn’t create a basket of uncorrelated cryptos… it doesn’t have any macros in there.”
He said judgment fills that gap.
“That’s where the human knowledge and experience comes in and the intuition… That the AI agent doesn’t have or the LLM.”
He also warned against treating AI as magic. The tool will not deliver “some sort of magic that creates an infinite money loop.” That caution matches the wider market, where few experts back hands-off trading bots.
Moreno said institutions trust data but keep testing it.
“They do trust it but they verify a lot, and are continuously monitoring if the data remains relevant.”
Inside the Models
Professional funds treat AI as infrastructure, not a crystal ball. Edwards built his firm around large, tested models.
“We build hundreds of metrics and we also use hundreds of other data sources to build out comprehensive models… Combining onchain technicals and macro data for many years to build out trading models.”
Capriole’s Macro Index reflects that approach. The firm combines more than 60 on-chain, macro, and equities metrics into one machine-learning model. Most data platforms publish thousands of metrics, yet models still need careful curation.
Cowen is building his own bot from the ground up.
“Right now all the bot does really is regurgitates things that I say. It’s almost like an AI version of me.”
He avoids training on low-quality AI output to prevent model decay.
“I don’t want it to use AI slop that’s out there to create more AI slop”
Van de Poppe runs his fund the same way. AI writes the base of his trading algorithms, but a human keeps steering it, or it keeps “working on stuff that is wrong for your system.”
The Data Behind the Models
Every model depends on the data beneath it. Moreno gave the sharpest example of a data edge.
“They will trade for example mining stocks instead of waiting for your quarterly report you’re tracking in real time actually what they’re mining.”
Network hashrate offers one such real-time signal. It tracks how much computing power miners commit to Bitcoin each day.
The same method applies to equity exchanges. Bitcoin miner stocks have drawn fresh attention as AI infrastructure spending climbs. Julio Moreno continues:
“Some of the crypto exchanges have also started trading on stock exchange and so you can be monitoring the trading volume to assess the revenues.”
Cowen added that data quality decides the outcome. He values records from before the AI era.
“Data before 2022 in some ways is actually really valuable because it was data before all the AI stuff was even here.”
For institutions and retail traders alike, the lesson holds. AI compresses the work and widens access, but the advantage flows to operators with clean data and the judgment to steer the model. As adoption spreads, that judgment becomes the real differentiator.
The post AI Won’t Make You a Good Trader, But Here’s How the Pros Use It Anyway appeared first on BeInCrypto.
