2024 is the year analytical and reasoning capabilities leapt forward. The generated images became jaw-droppingly realistic, and the videos unsettlingly lifelike. Yet according to Demis Hassabis, the CEO of DeepMind, “2025 will be the year of AI agents.” The game is changing: it’s no longer about generating content, but about acting autonomously.
You might already be using them without realizing it. Tools like ChatGPT’s “deep search,” Perplexity, or Gemini break down your question, scour dozens of web pages, and deliver — sometimes after a few minutes — a detailed response, almost like a mini research paper. These are agents.
But AI agents aren’t “just” AIs that answer questions. They are AIs that pursue a goal, make decisions, choose tools to use, and act, sometimes through multiple steps, without human intervention.
Still early days… but spectacular nonetheless
OpenAI launched Operator in January. It can plan trips or purchase products online from start to finish, completely autonomously. It’s available only to premium subscribers and remains low-key for now. But it was Manus, a Chinese AI agent unveiled in March, that sent real shockwaves through the field.
Manus can take control of a browser, navigate complex websites, fill out forms, connect to business tools — and even launch a design in Canva. Sure, the result isn’t very pretty yet. But when you watch an AI using software like a human would, one thing becomes clear: the threshold of autonomy has been crossed.
And given the speed at which AI evolves, we know these clumsy prototypes will soon become highly effective assistants for repetitive, precise, and structured tasks.
The Era of “Super-Agents” Is Approaching
Google is also developing agents, like Project Mariner, designed to act within Chrome. But again, it’s OpenAI making the biggest waves, with its announcement of a lineup of “super-agents”:
- Executive & Expert Agent: $2,000/month
- Developer Agent: $10,000/month
- PhD-Level Research Agent: $20,000/month, capable of formulating scientific hypotheses and coding autonomously.
We’re entering a whole new dimension.
AImpact on a communications Agency: A new way to work
As entrepreneurs in the communication field, we’re already feeling the shift. Writing a press release or social media post without AI now feels outdated. And tomorrow?
Tomorrow, we might delegate the entire chain: trend monitoring, editorial strategy, content creation (text, images, video), publishing schedules, community management, influencer targeting and activation, even email or SMS follow-ups.
An AI agent could track key conversations around your brand, detect weak signals, suggest a strategy, and implement it — from start to finish — without lifting a finger.
And if that sounds futuristic, remember this quote from Sam Altman, CEO of OpenAI:
“We’re betting on the first year a billion-dollar company will be launched by a single person.”
Entrepreneurs: Now it’s our move
While we’re convinced that humans will still play an essential role in leadership, supervision, and ethical decision-making, it would be dangerous to believe this revolution will leave us untouched. Homo homini lupus est. Every job, every role, every process must be reimagined through the lens of autonomous AI.
There’s no need to warn our readers of a dystopia à la Gattaca. Instead, we should see AI as a kind of “new light bulb”, like Edison’s, that is still diffusely illuminating our era.
But we must act now. Because for us, as entrepreneurs, this is not a threat — it’s a vast playground. A powerful lever for productivity, creativity, and optimization. As long as we don’t stay on the sidelines.
Now is the time to shape the agency (or company) of tomorrow. Now is the time to integrate these tools, reinvent our methods, and reaffirm what only humans can bring: vision, ethics, creativity, emotion.
AI agents won’t replace us. But they do force us to transform.
Op-ed by Edouard Fillias and Alexandre Villeneuve, CEO & Co-founders of JIN.
