The rise of AI traders – or why portfolio managers can no longer go it alone

Ask any successful portfolio manager about the first thing they do on waking up and they’ll say read the speeds and feeds from the market. Wake up, find out what’s going on, work out the strategy for the day – or at least the next hour or two.

In our grandparents’ day, traders picked up telegrams and read the financial papers. Now there are thousands of sources that must be collated and understood before trades, Bloomberg terminals, custodian filings, blockchain data, over-the-counter desk spreadsheets, the digitization of real-world assets (RWA), and brokerage feeds, all need to be taken into account. And quickly.

Data friction from marrying all these disparate sources together slows trading and reactions. By the time an analyst has scanned the available information a trade may be behind the market reaction, or worse, actively unhelpful. Getting information flows running smoothly is a priority.

Fragmented data hides a whole host of very serious problems, particularly with exposure to risk – both financial and regulatory. Working on off-chain prices misses the dynamics playing out on decentralized exchanges. Errors that would surface immediately in a unified system can hide in the gaps between siloed databases, leaving a broker open to regulators and market losses.

These pressures are only going to get more problematic as traditional financial institutions move into the blockchain environment. The increasing mix of RWA and digitization has opened up whole new avenues of arbitrage and trading. Failing to recognize and act on this would be a potentially fatal mistake.

In with the old, and the new

TradFi and DeFi markets are merging, and old-school traders who initially shunned digitized assets now realize that they’re missing out. There’s no room for ignoring these bountiful new markets and data friction is slowing their adoption among traders who rely on traditional techniques.

The fact of the matter is that cryptocurrencies, blockchain, and digital assets are now the norm for online commerce and the market must recognize that. Failure to do so will lead to direct losses and indirect losses – the loss of huge potential profits. The days of online currencies being something interesting but unimportant are long gone.

To complicate the situation further, AI agents are proving faster than their fleshy creators in analyzing markets and making buying recommendations. A human hand is still needed on the tiller but it’s clear that software is overtaking humanware in terms of data analysis, be it market maneuverings or basic wallet transactions.

A 1979 IBM manual warned that “A computer can never be held accountable, therefore a computer must never make a management decision.” This was true at the time, and remains a core of how responsible companies operate, but the fact of the matter is that there are too many data feeds for a human to manage personally. It’s time to get the professionals in.

Crypto Flash Crash: $3.21B gone in 60 Seconds

On October 10, 2025, $3.21 billion in crypto positions vanished in 60 seconds, part of a $9.89 billion liquidation cascade driven by extreme leverage and collapsing liquidity. This report breaks down the crash’s mechanics with data-driven insights for traders and analysts.

Picking out the threads of value

This is where digital asset platforms step in. They take this data and put it into a format that brokers can react to, complete with oversight to give what’s called an Alpha edge on trades. AI agents can amalgamate data from a multiplicity of feeds, but human operators need to read this and make decisions.

Traders need a framework to ask the right questions and get the information they need. With millions of trades a minute, the information overload is too much for a single person to handle, so AI will take up some of the workload, once properly guided.

AI agents can stress-test multiple scenarios far more quickly than a human operator can, just as spreadsheets last century relieved accountants from endless plotting of financial data. An analyst can game-play multiple scenarios by weaving in data from a huge number of inputs, and potentially spot erroneous trades and issue warnings of possible dangers ahead. And with the new generation of tools, traders don’t need to be computer whizz kids to navigate SQL queries, construct API calls, and merge the data into a cohesive whole.

Not to say human knowledge is automatically discarded. A smart business intelligence platform takes its prompts not only from digital data, but also from human analytics and papers published by specialists in the field. After all, AI has certain limitations in this area. It lacks the left-field thinking from humans that can reveal insights into market trends. That data needs to be added in also.

It’s a conjoined future

As we’ve seen in the past six months, digital trades are subject to wild fluctuations in price and portability. But these can be mitigated by bringing together fragmented data and real-time interpretation.

Just as important: recognizing what data is real and actionable, and what may be mistaken indicators. Poorly interpreted information can be just as dangerous as a call not made based on solid groundings, and meeting regulatory requirements.

So an analyst’s platform needs to sort out the wheat from the chaff, give actionable advice, and mix in traditional data sources from those taken from new platforms. A unified approach by specialists in data analytics allows AI agents to monitor markets and make predictive suggestions that the trader can act upon based on their market experience and the information they are receiving.

In a world where split-second decisions can count for millions on trades, it’s clear that a human operator can’t keep up on their own. Nor can an AI agent make potentially financially crippling decisions without human oversight. What’s needed is a melding of the two, to make the best decisions possible in an increasingly last-minute market.