Can We Just Talk About AI for a Sec? (Stepping Back to Actually Understand What’s Going On Here)

Look, we get it.

Your LinkedIn feed is 90% AI posts right now. Every company is suddenly “powered by AI,” “redefining AI,” or “leading the AI revolution.”

At this point, it’s starting to feel like the early 2000s when every cereal brand added “whole grain” to their boxes.

So yeah, AI is everywhere. But under all that noise, there are real differences between the types of AI people keep throwing around: predictive, generative, and now, “agentic.”

And if you’ve ever nodded along in a meeting while someone said “agentic AI” and silently hoped no one would ask for your take, this one’s for you.

We’re going to break it down like normal humans. No PhD required, no buzzwords allowed. Just a clear, honest look at what each type actually does, why it matters, and how to tell the difference when everyone online sounds like a TED Talk.

Predictive AI: The Crystal Ball (Sort Of)

Predictive AI is the most practical, least flashy member of the AI family. It looks at data from the past and says, “Hey, this pattern usually means X is going to happen next.”

It’s the engine behind your Netflix recommendations, your credit card fraud alerts, and your sales forecasts. It’s not magic, it’s math. But it’s good math.

How it works: Predictive models use historical data to identify patterns, then extrapolate them into the future. It’s all about probabilities, not guarantees.

Use cases:

• Forecasting demand or spend
• Predicting supplier risk or delivery delays
• Spotting anomalies (aka “something looks weird here”)

It’s kind of like that friend who tells you, “Don’t text them, they’re going to ghost you.” Not because they’re psychic, they’ve just seen this before.

Generative AI: The Creative One

If Predictive AI is the math geek, Generative AI is the art major with a side hustle. It creates new content—text, images, code, videos—based on patterns it’s learned from existing data.

ChatGPT, DALL·E, Midjourney all live here. They don’t “know” things; they remix what they’ve seen in statistically plausible ways.

How it works: Trained on huge datasets, these models learn relationships between words, pixels, or sounds. When prompted, they generate new stuff that feels original, though it’s all inspired by the past.

Use cases:

• Writing RFP drafts or supplier outreach emails
• Summarizing massive spend datasets
• Brainstorming creative solutions or documentation

In other words, Generative AI is your overachieving cousin who paints, codes, and writes poetry. Brilliant? Yes. Reliable with numbers? …eh.

Agentic AI: The Grown-Up in the Room

Here’s where things get spicy. Agentic AI is about doing, not just about predicting or creating.

This is AI that can take a goal, plan the steps, execute them, and learn from the results. Rather than simply making suggestions, it takes action and follows through. It’s the bridge between automation and autonomy.

How it works: Agentic systems combine reasoning, memory, and planning. They can break down tasks, make decisions based on context, and adjust as they go.

Use cases:

• Procurement assistants that find suppliers, negotiate, and finalize purchases autonomously
• AI agents that monitor inventory, trigger workflows, or optimize spend without needing human nudges
• Any system that says, “I’ve got this,” and actually means it

Think of Agentic AI as the intern who became your project manager.  It goes beyond forecasting what might happen and actually takes steps to make it happen.

And this is where we’re heading: a world where AI isn’t a tool you query, but a teammate you trust to handle the boring stuff (so you can focus on the strategic stuff).

What Agentic AI Looks Like at Fairmarkit

At Fairmarkit, we’re bringing this “grown-up” version of AI to procurement: where repetitive work, complex data, and human judgment collide daily.

Traditional procurement automation has always stopped at suggesting next steps: a list of suppliers, a pricing analysis, maybe a recommended email draft. With agentic AI, the process evolves from generating insights to actually executing on them, transforming recommendations into real outcomes.

Fairmarkit’s agentic AI can autonomously manage sourcing workflows from start to finish: it identifies qualified suppliers, drafts outreach messages, negotiates pricing, and even initiates purchases, all while keeping humans in the loop for oversight and strategy.

It’s like having a digital teammate who handles the tactical grind 24/7, freeing procurement teams to focus on what really matters: building relationships, driving innovation, and making strategic calls that AI can’t (and shouldn’t) replace.

In other words: it goes beyond smarter automation, delivering truly intelligent and decisive action.

The “Etc.” Category

Not every AI fits neatly into these buckets, but a few other types deserve honorable mentions:

• Reactive AI: The OGs of the AI world. Think of Deep Blue (the chess-playing computer). They react to inputs but don’t learn from them.
• Cognitive AI: Mimics human-like reasoning, often used in simulations or diagnostics. More “thinking,” less “acting.”
• Hybrid AI: The smoothie of AI. Mixes predictive, generative, and agentic capabilities into one model for complex applications.

These are the supporting cast. Predictive, Generative, and Agentic get the spotlight, but the others still keep the show running.

Why This Matters (The Human Take)

Because not all AI is created equal. Predictive helps you see ahead. Generative helps you create. Agentic helps you get things done.

Understanding which is which helps you cut through the hype and focus on what actually drives outcomes. Especially in enterprise functions like procurement, where the difference between “smart insights” and “self-executing workflows” is massive.

At the end of the day, AI isn’t a mysterious black box but rather a collection of really fancy math tricks. But when that math is designed to make life easier for actual people, that’s when the magic happens.

So next time someone drops “agentic AI” in conversation, you can smile, sip your coffee, and say:

“Ah, the one that actually does stuff.”