In just one year, humanity has achieved the impossible: we have renamed basic statistics “artificial intelligence” and transformed simple to-do lists into “multimodal autonomous agents.” This is the story of how Silicon Valley marketing departments (and others!) fooled us all—and continue to do so with ever-increasing creativity.

In 2024, any spreadsheet formula with a modicum of vitality was hailed as revolutionary AI.

In 2025, we moved on to calling automated workflows “autonomous digital employees with reasoning capabilities.” Let's look at how we got here, with a healthy dose of sarcasm and a growing sense of existential angst.

In 2024, the tech industry collectively decided that if a piece of software could count to ten without using its fingers, it deserved the “AI” label. Startups raised billions (no longer just millions) by slapping neural network stickers on algorithms that, at their core, were traditional statistical models—techniques older than your great-grandmother's panettone recipe.

Venture capitalists continued to fall in love with pitches that included phrases like “powered by generative artificial intelligence,” even when the product was essentially an if-then rule system with a fancy interface.

The reality? While large language models (LLMs) have made impressive strides in generating coherent text, AI continues to struggle with fundamental problems of reasoning, contextual understanding, and hallucinations. Generative chatbots produce responses that sound plausible but contain factual errors.

Meanwhile, marketing departments have started putting “AI” in every product possible: from the “AI-Enhanced Refrigerator” that reminds you when you're out of milk to the “Smart AI Toothbrush” that analyzes your brushing patterns with “deep learning algorithms” (read: a timer with statistics).

Take AI-generated humor, for example. Progress has been real, but the results are still far from truly funny. Modern language models can now understand some aspects of humor, but their original jokes tend to be predictable or just plain weird.

Take the case of AI-generated humor. Progress has been real, but the results are still far from truly funny. Modern language models can now understand some aspects of humor, but their original jokes tend to be predictable or just plain weird.

If an LLM attempted stand-up comedy, we might hear jokes like: 'I asked my AI assistant to make dinner, and it ordered takeout... of my hard drive. It was cooked to byte perfection!' — the kind of humor that would make 'knock knock' jokes seem like high literature.

In 2025, the automation tools industry has completed its transformation. Platforms such as Zapier, Make, and n8n no longer sell simple “automation” — they now offer “autonomous agent ecosystems” and “neural automation networks.”

The main difference? According to experts, ‘AI agents’ differ from traditional automation because they can ‘process multimodal information,’ ‘show reasoning,’ and ”think proactively.” In reality, as highlighted in Agenda Digitale, AI agents are simply “software components that interact with an enhanced LLM and perform actions with the outside world,” often using the same conditional logic as traditional automation, but with more elegant terms.

Zapier now boasts “Proactive Agents that anticipate your business needs” (read: scheduled automations that activate at set times), while n8n offers “Multi-Agent Orchestration with Contextual Memory” (i.e., workflows that can access previous data).

Under the surface, the code still looks like this:

Yet, according to press releases from 2025, this is “Cognitive Hyperautomation that implements causal reasoning through synergized business ontologies” — a fancy way of saying “it does different things based on conditions.”

The real novelty of 2025 is the rise of “multimodal agents” — systems that can (in theory) work with different types of input: text, images, audio. Automation platforms now promise that their agents can “analyze photographs of your documents, extract data, and automatically fill out forms” — a feature previously known as “OCR with form processing,” a technology available since the 1990s.

Make now advertises “Empathic Agents” that can “perceive the emotional tone of emails and respond appropriately” — translatable as “a sentiment classifier that categorizes text as positive or negative.”

According to IBM, the official definition of an AI agent is “a system or program that can autonomously perform tasks on behalf of a user or another system by designing its own workflow and using available tools.” The main difference from traditional chatbots is that AI agents can operate without continuous user input, make decisions, and plan future actions.

However, as IBM experts pointed out in a recent article, “it's impossible to take two steps in the tech media landscape without stumbling across an article hailing 2025 as the year of the AI agent,” but the reality is much more nuanced. Marina Danilevsky, Senior Research Scientist, commented that the adoption of agents in the workplace “will have growing pains” and that there will always be cases where “as soon as something becomes more complex, you will need a human.”

Looking toward the end of 2025 and the beginning of 2026, the future is clear: expect “Meta-Cognitive Agents” that promise to “think your thoughts” (read: record logs of their own operations), “Quantum Agents” that can “exist in overlapping decision states” (translation: they have randomized functions), and “Empathic Digital Twins” that “develop an emotional relationship with your data” (i.e., they display different emojis based on trends in graphs).

According to a recent report by Unite.ai, the AI agent market is growing at an extraordinary rate, with projections indicating growth from around $5 billion in 2024 to over $47 billion by 2030. However, the validity of these numbers is as questionable as the actual ability of agents to deliver on their creators' promises.

The lesson remains the same: never underestimate the tech industry's ability to repackage old ideas as revolutionary. As one Reddit user aptly observed: “Using an ‘Autonomous AI Agent’ in 2025 is like hiring an assistant who, for every request, asks you exactly what to do, how to do it, when to do it, and then claims to have figured it out on their own.”

And if that's not the future of AI, what is?

For sources specifically about automation, see:

Welcome to Electe’s Newsletter - English

This newsletter explores the fascinating world of how companies are using AI to change the way they work. It shares interesting stories and discoveries about artificial intelligence in business - like how companies are using AI to make smarter decisions, what new AI tools are emerging, and how these changes affect our everyday lives.

You don't need to be a tech expert to enjoy it - it's written for anyone curious about how AI is shaping the future of business and work. Whether you're interested in learning about the latest AI breakthroughs, understanding how companies are becoming more innovative, or just want to stay informed about tech trends, this newsletter breaks it all down in an engaging, easy-to-understand way.

It's like having a friendly guide who keeps you in the loop about the most interesting developments in business technology, without getting too technical or complicated

Subscribe to get full access to the newsletter and publication archives.