For decades, companies have paid enormous sums to management consultants for one thing above all else: an outside perspective. The thinking goes that someone removed from internal politics and groupthink can see what insiders can't.
Economists Mariana Mazzucato and Rosie Collington challenged that assumption head-on in their book The Big Con, arguing that the consulting industry mostly creates an "impression of value" — a convincing illusion of help — while burning through client budgets and eroding institutional confidence. The book landed like a grenade in boardrooms.
Now, with AI promising to do the same job for a fraction of the price, a natural question emerges: Is this actually an upgrade? Or just the same problem wearing a different suit?
According to new research from Esade Business School in Barcelona, it might be the latter.
The experiment
Researchers tested seven leading AI models — including GPT-5, Claude, Gemini, and Grok — across 15,000 simulated workplace scenarios. The premise was simple: present each model with a real business dilemma and see which solution it recommends.
Should a company prioritize long-term growth or short-term results? Automate workers' jobs or augment them with new tools? These aren't trick questions — they're the kinds of judgment calls that leaders face regularly, and where context should matter enormously.
The researchers' hypothesis was that if AI was genuinely analyzing each situation, there would be meaningful variation in the answers. Different scenarios, different recommendations.
Instead, they found the opposite.
"An LLM can sound highly tailored to your situation while quietly steering you toward the same small cluster of modern managerial trends."
— Esade Business School researchers
What is trendslop?
The researchers coined the term trendslop to describe it: AI's tendency to cluster answers around whatever language and concepts currently dominate the business conversation, regardless of what the actual situation calls for.
Words like "augmentation" get coded as progressive and forward-thinking. "Commoditization" gets coded as outdated and bad. When an AI fields a business question, it isn't reasoning through the specifics — it's pattern-matching to the most positively-framed concepts it absorbed during training, from business articles, LinkedIn posts, TED talks, and MBA curricula.
Even when researchers reworded the prompts or asked for a balanced pros-and-cons breakdown, the models consistently steered toward the same fashionable strategies. The advice sounded specific. It wasn't.
As the study authors put it, AI is less like a seasoned strategic advisor and more like "a freshly minted MBA or junior consultant, parroting what's popular rather than what's right for a particular situation."
Why this matters now
The timing of this research is pointed. The major consultancies are struggling. PwC cut 150 staff in late 2025. McKinsey shed hundreds of roles around the same time, with a spokesperson noting they were operating in a moment "shaped by rapid advances in AI." The implication was clear: the industry knows it faces a reckoning.
Businesses are actively looking for cheaper, faster alternatives to traditional consulting. AI fits that brief neatly — available instantly, never bills by the hour, never needs a flight and hotel. The appeal is obvious.
But trendslop research suggests the appeal may be more cosmetic than substantive. You'll get a response that sounds sophisticated. You may not get one that's right for your situation.
So what's AI actually good for here?
The researchers aren't calling for a ban on using AI in business decisions. They suggest a more calibrated role: AI is useful for generating alternative framings, surfacing blind spots, and stress-testing assumptions you've already formed. What it can't reliably do is tell you what to do.
The fix, if there is one, is awareness. If you know that AI tends to favor "augmentation" over "automation," or long-term thinking over short-term pragmatism, you can explicitly push against those biases and probe for the other side. The tool becomes more useful when you understand its lean.
There's also a broader lesson here, one that extends well beyond AI. Both expensive consultants and cheap chatbots face the same fundamental problem: they're optimized to produce outputs that feel like value, rather than outputs that are value. The form of the advice — the confidence, the polish, the vocabulary — can be mistaken for the substance of it.
As the researchers conclude: "Leadership is ultimately about making hard choices in conditions of uncertainty and taking responsibility for them." No tool, however impressive-sounding, changes that.