The Hype!
The hype around AI these days can feel deafening. Every headline screams about breakthroughs, tools, and risks. Honestly, I’ve felt the same sense of overwhelm, especially now about agentic AI.
To get some clarity, I turned to a book by a team of seasoned AI consultants: Agentic Artificial Intelligence: Harnessing AI Agents to Reinvent Business, Work, and Life. It’s worth a read if you want a high-level framework about agentic AI, some use cases, and insights on its potential impact on business and society.
From Tools to Collaborators
We’ve all spent the last couple of years (has it really only been that long?) talking about generative AI — systems that produce text, images, or code in response to prompts. But this book makes the case that something bigger is on the horizon.
Agentic AI isn’t just about one-shot outputs. It’s about continuous, goal-driven systems that reason, plan, and adapt until the job is done. In many ways, these agents look closer to what the public always imagined “real AI” would be.
The authors draw a powerful distinction between tools and collaborators. A tool gives you a hammer when you ask for one. A collaborator asks whether a hammer is really what you need, remembers how you used it last time, and then helps you build something better than you could have managed alone. That shift, from tool to collaborator, is the heart of the agentic story.
Why the “Agentic Loop” Matters
One concept that stuck with me is the agentic loop: memory, planning, and reflection. An agent that remembers past interactions, plans multiple steps ahead, and reflects on what worked (or didn’t) starts to feel almost human. That’s no longer simple automation. That’s a system that grows smarter with use.
The book shines when it gets practical. In business, agents can become digital co-workers, taking on repetitive, low-value tasks so people can focus on judgment, relationships, and creativity. In personal life, they might act as tutors, assistants, or even lifelong learning companions.
My Blog as a Test Case
Toward the end, the book suggests building a personal agent stack. So I thought: what if I used this blog as my test case?
Right now, I do a lot of manual work — re-tagging older posts, tweaking for SEO (rarely), building an editorial schedule, and brainstorming engagement strategies. It’s repetitive and sometimes draining. An agentic system could not only automate those steps, but also continuously monitor analytics, adjust my strategy, and even suggest pivots when reader interest shifts.
In other words, it could be the collaborator that helps me grow this blog with less grunt work and lets me focus more on creativity.
A Small Experiment for You
If you’re curious about agentic AI, you don’t need a whole stack to start. Here are a few light experiments you could try this week:
- Automate a task you usually do manually. For example, ask an AI to summarize your last 10 emails into 3 bullet points every Friday and send to your inbox. Notice what changes in how you work.
- Add “memory” to your AI use. Keep a running log of prompts and answers in a doc, or by leveraging Notion, and revisit them. Over time, you’ll see how your AI interactions “stack” together.
- Designate an AI “role.” Try treating AI as a teammate — e.g., your research assistant or your writing coach — and see how shifting that frame changes the interaction.
These are small steps, but they give you a feel for how agents differ from simple one-off tools.
The Big Picture
The truth is that rolling out an agentic system in the enterprise is still tough — the failure rate is around 95% so far. But that doesn’t mean it won’t happen. Transformations in technology often look impossible right up until they’re inevitable.
So for me, experimenting with agentic AI isn’t just about efficiency. It’s about practicing now for a future where working with AI collaborators will be as normal as opening your laptop.
Where could an AI collaborator step in to make your work (or life) lighter this week?
Meet John Song 