The headline screamed: 'AI Trading Now 55% of Exchange Volume.' The token pumped 12% in an hour. But when I dug into the on-chain data, I found that the exchange's total volume had actually shrunk by 30% over the same period. The 'AI' share was rising because everything else was collapsing. This is the same trap that ensnared Crypto Briefing when it reported on ISG's claim that AI sales exceeded half of total revenue. The pattern is universal: in a bear market, percentages become weapons of mass deception.
ISG (Information Services Group) is a Nasdaq-listed technology research and advisory firm. Its core business is helping enterprises select IT vendors and benchmark tech spending. In recent years, as AI has swept through corporate budgets, ISG's advisory services have benefited. But the claim that 'AI sales exceed half of revenue' is a statement about revenue composition, not revenue growth. And that distinction is everything. Where liquidity hides, narrative finds its voice β and right now, the AI narrative is shouting from the rooftops.
The consulting industry is cyclical. In 2024β2025, traditional IT outsourcing and ERP consulting budgets have been squeezed. Clients are reallocating spending toward AI pilots β small, short-term projects that often yield uncertain returns. So when ISG says AI now constitutes 50%+ of revenue, it could simply mean that traditional revenue has collapsed, not that AI revenue has surged. Without total revenue growth, the 'majority' claim is a red flag, not a victory lap. This pattern is not unique to traditional consulting. In crypto, we see similar 'majority' claims everywhere: 'AI-driven trading now accounts for 60% of our volume,' 'DeFi yields now represent 70% of our revenue,' 'Layer2 transactions now exceed 80% of all activity.' These numbers are often presented as proof of adoption, but they hide the same denominator question. In a bear market, when overall volumes and TVL have shrunk, a rising share can be a sign of decay, not growth.
Let's break down the ISG claim using the framework I apply to any crypto metric. First, the numerator: what counts as 'AI sales'? ISG could be counting pure AI project delivery, digital transformation contracts with any AI component, AI vendor selection advisory, or research subscriptions that include AI topics. Each definition yields a vastly different number. In crypto, the same ambiguity exists. When an exchange claims 'AI trading volume,' does it mean trades executed by algorithms? Or trades initiated by users using AI tools? Or just any trade on a pair that includes an AI-related token? The lack of definition is a hallmark of narrative-driven reporting.
Second, the denominator: is total revenue growing or shrinking? For ISG, we don't know. The article didn't mention it. In crypto, the denominator is often the total market volume or TVL. If overall volume is down 40% year-over-year, an AI-driven volume share of 60% could mean AI volume has stayed flat while everything else collapsed. That's not adoption; that's a shrinking pie. I learned this the hard way during the 2020 DeFi Summer, when I joined a DAO building a cross-chain bridge. We proudly announced 'TVL exceeding $100M' without disclosing that 80% of it was our own governance token. The lesson: always trace the source. If it's the company's own IR, treat it as a lead, not a fact.
Third, the source structure. The ISG story originated from a company statement (IR narrative) and was relayed by Crypto Briefing with no third-party verification, no earnings call quotes, no analyst commentary. In crypto, we often see 'research' reports from protocols themselves, repackaged by media outlets as news. The result is a feedback loop of self-serving narratives. The more a narrative emphasizes 'low risk,' the more leverage is hidden underneath. I spent weeks mapping the balance sheet overlap between Celsius and Genesis after the Terra collapse, and the pattern was clear: the illusion of control in a fluid world is that you think you're buying exposure when you're actually buying narrative.
Fourth, the 'low-risk investment entry' claim. The Crypto Briefing article suggested that ISG offers a low-risk way to invest in AI. This is a logical fallacy. ISG's revenue is service fees, not AI operating leverage. Its earnings volatility is far lower than chipmakers or model developers, but it also offers no valuation premium. In crypto, we see the same fallacy: buying a token because it has 'AI' in its name does not give you exposure to AI; it gives you exposure to a token's liquidity dynamics. During the 2023 AI hype cycle, I watched countless 'AI coins' pump 10x, then collapse 90% as the underlying projects had no revenue. Chasing ghosts in the algorithmic machine is a dangerous game.
Now let's apply this to a real crypto example. Consider a mid-cap exchange that recently announced 'AI-powered trading now represents 55% of total volume.' The market pumped the token 12%. But looking at the on-chain data, I found that total volume had fallen 30% quarter-over-quarter. The exchange's overall trading volume was shrinking. The 'AI' trades were largely small retail orders routed through a third-party API. So the 55% share was actually a sign of decay: traditional market makers had left, leaving only low-quality flow. The token's price is now 40% below where it was at the announcement. This is the same denominator trap.
Another example: a DeFi protocol announces that 'AI-enhanced yields now account for 60% of TVL.' But the TVL is inflated by a liquidity mining program that pays 50% APY in the protocol's own token. The 'AI' part is just a rebranding of a yield aggregator. When the incentives dry up, TVL will collapse. Yield is often a function of liquidity incentives, not protocol utility. This is a signature observation from my days analyzing Curve emissions. The yield trap is set by the ease of entry.
The most dangerous part of the ISG story is the 'low-risk investment entry' label. It's not just wrong; it's a trap. In crypto, we see this constantly: 'Buy this token for safe exposure to DeFi,' 'This NFT is a low-risk collectible,' 'This Layer2 is a safe bet.' The reality is that all these assets are high-risk, high-volatility, and subject to systemic contagion. The Terra collapse taught me that hidden leverage can turn a 'safe' yield farm into a death spiral. I spent weeks mapping the balance sheet overlap between Celsius and Genesis, and the pattern was clear: the more a narrative emphasized 'low risk,' the more leverage was hidden underneath.
The ISG claim also ignores the biggest risk: AI is cannibalizing the consulting business model. As AI tools reduce the need for junior analysts, consulting firms will face pricing pressure. ISG's 'AI revenue' might be growing, but its traditional high-margin business is eroding. In crypto, the same dynamic plays out: AI trading bots are reducing the need for human traders, which could eventually erode exchange fee revenue. The protocols that embrace AI may be sowing the seeds of their own margin compression. The trap is in the ease of entry.
Moreover, the media outlet's pivot is itself a signal. Crypto Briefing, a crypto media outlet, is now writing about AI consulting firms. Why? Because crypto ad revenue has dried up in the bear market, and AI is where the advertising budgets are. This is a classic case of media economics driving content. As a reader, you should ask: what is the incentive behind this story? Is it to inform you, or to generate clicks? Chasing ghosts in the algorithmic machine means recognizing that the narrative is often manufactured.
So what should we take away from the ISG story? First, treat 'majority' claims as suspect until you see the absolute numbers and the denominator. In crypto, this means demanding total revenue, total volume, and total TVL before celebrating any share metric. Second, verify the source: if it's a company's own IR, it's a lead, not a fact. Third, beware of 'low-risk' narratives in a high-risk industry. The only low-risk position is cash β and even that loses value to inflation.
As we navigate this bear market, the key is survival, not gains. Use data to judge which protocols are bleeding, not which are pumping. Where liquidity hides, narrative finds its voice β but your job is to find the liquidity, not the narrative. Volatility is just information wearing a mask; read the data, not the headlines. When the next 'AI majority' claim hits your feed, will you ask for the denominator, or will you chase the ghost?