The AI Agentic Revolution… and the end of work as we know it

Frederic Etiemble
May 15
6 min read

Updated: Aug 12

Every once in a while, the rules of work change. Quietly at first. Then suddenly.

If you’ve felt like your habits, tools, or even your role are being redefined lately, you’re not imagining it. If you’ve felt that the way you work differs from some of your colleagues, you’re not alone.

We’re currently living through the AI Agentic Revolution, a shift from doing the work to briefing the AI agents that do it for you.

This shift occurs in five steps, creating five distinct worlds of work. It is unlikely that we are all in the same one, hence the significant differences we've noticed lately in how people approach their work.

Let me guide you through it. Here’s the map:

Five worlds. Five rules of work. Which one are you operating in?

1. The Old World

When you have something to do, the general rule is “Figure It Out".

Before ChatGPT, let’s call this period BC (Before ChatGPT), our work approach was simple. You had a task, so you figured it out. Whether you already knew how to handle it, searched on Google, or asked a colleague, the work always came back to you.

Tools were informational, not operational. They helped you learn but not do the work itself. Even when searching for something, you had to click, read, compare, decide, and finally execute the task.

This was - and still is for some of us - the era of the DIY worker. Knowledge search engines and collective resources supported the human worker, but ultimately, you were at the center of the effort.

In that world, you have to know why you need to complete the task, how to do it, and most importantly, you still have to do it.

2. The World of AI Assistants (2022)

When you have something to do, the rule becomes Ask ChatGPT.

Then came the disruption in late 2022 with the release of ChatGPT.

Suddenly, instead of merely searching, we asked. ChatGPT didn’t just tell us how to do something, it often just did it for us, e.g.:

Write an email.
Summarise a paper.
Translate a paragraph.
Fix a piece of code.

We all gained an assistant, marking the beginning of a new behaviour. You had a task? You simply asked the AI Assistant to handle it, and often, that was enough.

Millions of knowledge workers shifted overnight from doing the task to typing a prompt.

Of course, the AI Assistant could not perform every task we faced daily. However, when it could not complete the task directly, it could provide a roadmap on how to do it.

As a result, search engines began to feel slower, older, and less helpful. Assistants became ubiquitous with ChatGPT joined by Claude, Perplexity, Copilot, Gemini, etc. The era of the assisted worker had begun.

In that world, you need to understand why you require to complete a task, but you don't necessarily need to know how to do it, the assistant knows. But often, you still have to execute the task in the end.

3. The World of AI Agents (2024)

When you have something to do, you find the right AI tool for the job.

By 2024, the rules had evolved again. Instead of asking one general-purpose AI assistant to handle everything, we started to delegate tasks to specialised AI tools, e.g.:

A design tool to build our PowerPoint decks.
A research tool to analyse market data.
A meeting tool to transcribe and synthesise our meetings.
A video tool to create our videos.

There’s an AI for that (TAAFT) is the Google of that world. It is a platform dedicated to connecting users with artificial intelligence tools and resources. At the time of this writing, there are close to 40,000 specialised AI tools referenced on TAAFT.

They’ve given birth to the realm of the enhanced worker.

In that world, you need to understand why you require to complete a task, but you don't necessarily need to know how to do it. You can delegate execution to an AI tool, but always with set guardrails. AI tools are enhancing our productivity to levels previously unseen, but we always keep full control of what to do with the output.

4. The World of AI Agents (Now)

When you have something to do, now you just ask an agent to do it.

A few months later, the rules governing how we work have changed again with the rise of autonomous agents.

No longer do you need to select among the many AI tools, guide the work step by step, making the output of one AI specialised tool the input for the next one. Instead, brief your AI agent, give it a clear goal and let it work for you autonomously. We are no longer talking about a productivity tool, this is invisible labour.

Imagine asking your agent in ChatGPT to plan a 5-day work trip to Queenstown, New Zealand within a $3,000 budget, keeping two days open for sightseeing.

Provided with the right access, your agent would be able to:

Search and book flights and accommodation.
Schedule work meetings and add them to your calendar.
Order airport transfers.
Find dinner reservations and send confirmations.

Or you might brief Manus on a proposal to support a new product development for a client in the energy sector. Your Manus agent would:

Pull in and analyse past proposals.
Analyse available information on the client.
Ask clarifying questions.
Build the proposal and ask for your feedback.
Iterate the proposal with your feedback.
Create a video pitch of the proposal using your digital avatar.
Send the proposal to your client.

This is no longer a tool you control. It’s an agent you empower and that interacts with the world on your behalf.

Of course, autonomy introduces new concerns.

Are you comfortable giving your credit card to an agent? Allowing it to access your inbox or make booking decisions?

But once you start, you realise the real shift: your role is no longer to do the work. It’s to design the system of agents that accomplishes the work.

The world of the agent’s leader is starting. In this world, you still need to know why you need a task completed, but you no longer need to know what tool to use, how to accomplish it or even do it yourself.

5. The World of Connected Agents (Very Soon)

Soon you'll orchestrate teams of agents working together to achieve complex objectives.

While individual AI agents are powerful, the true revolution happens when they work together. This isn't a distant future - the foundations are being laid today through developments in agent-to-agent communication protocols and cross-organisational AI standards.

In this emerging world, you'll shift from managing individual agents to orchestrating interconnected teams of specialised AI agents that collaborate across departmental and organisational boundaries.

Imagine having to organise an offsite for the top 200 leaders in your organisation. Simply set the vision, clarify the objective and brief your AI Chief of Staff. It will then activate and coordinate all necessary agents.

This will include agents in your organisation, for instance:

An experience design agent working with the senior executive team on agenda and overall experience.
A communications agent managing pre-event engagement and materials.
A finance agent managing budgets and expense tracking.
An HR agent coordinating with regional offices on attendance.
An analytics agent gathering feedback and measuring event impact.

This will also include coordinating with agents outside of your organisation, including:

Agents from various hotels and venues the event will use.
Agents from airlines and other transport companies for flights and ground transportation for all attendees.
Agents from external speakers that will deliver keynotes during the event.
Logistics agents from all organisations involved in catering, AV, and room arrangements in selected venue.

Like a seasoned event management team, these agents collaborate seamlessly, share context across organisational boundaries, resolve conflicts, and deliver a flawless experience.

Your role? Define the vision, brief your AI Chief of Staff, be available to answer questions, provide feedback and… enjoy the conference.

No endless email chains. No timezone-juggling meetings. No interdepartmental friction over who’s doing what. No coordination headaches. No overlooked cultural sensitivities. No last-minute logistical crises.

In this world, you won't need to know why tasks need doing, who they involve, and how to complete them. Just establish and communicate the overall objective while your agentic workforce determines and manages the required tasks across organisational boundaries.

This is not assistance. This is orchestration. This is the future of work.

Which world are you working in?

Are you still Googling everything? Still trying to figure it all out on your own? Still juggling every task instead of allowing a team to support you?

While you’re busy doing that, others are:

Delegating to invisible teams of AI agents.
Designing workflows that operate 24/7.
Reclaiming time to focus on strategy, leadership, and creativity.

The AI Agentic Revolution isn’t coming; it’s already here. It marks the end of work as we know it, whether we like it or not.

The only question is: which world of work are you currently residing in?

Never before has William Gibson's quote "The future is already here, it's just not evenly distributed" felt so acutely true.

Reach out if you’d like to discuss our experiences of navigating these new worlds, or share some of your insights. We would love to hear from you!

About Fred

Executive advisor on strategy and innovation. Co-author of The Invincible Company, a guide to building resilience in organizations through corporate innovation. The book was shortlisted for the Thinkers50 Strategy Award in 2021.