Millie Marconi Profile picture
Founder backed by VC, building AI-driven tech without a technical background. In the chaos of a startup pivot- learning, evolving, and embracing change.

Aug 20, 12 tweets

If you’re building or investing in AI and don’t understand agents… you’re flying blind.

Here’s your shortcut: 10 core concepts every founder should know:

1/ Agentic AI

This is AI that doesn’t just answer questions it gets shit done.

Basically, It can plan, make decisions, and act without you babysitting it.

Think of the difference between asking a human for advice…

And having someone who actually takes the action for you.

2/ Agent

The basic unit of Agentic AI.

An agent is a piece of software that can see what’s going on, think about it, and do something to reach a goal.

Example: an ecommerce agent notices a product is almost sold out, checks sales data, and automatically places a reorder.

3/ Perception

How an agent “sees” the world.

It might read text, scan a video feed, listen to audio, or process data in a spreadsheet.

The more senses it has, the better it can understand its environment.

Without perception, the agent is basically guessing.

4/ Reasoning

This is the thinking step.

once it knows what’s happening, the agent figures out the best move.

Example: traffic is blocked → check historical traffic data → cross-reference with live feeds → decide on the fastest alternate route.

Reasoning turns raw input into smart action.

5/ Action

The moment the agent actually does something.

it might update a database, trigger a process, send a message, or even control a physical device.

reasoning is theory.
action is impact.

6/ Tool use

Agents don’t just use built-in features they can call apis, run code, search the web, or access databases.

This lets them solve problems far outside their “native” abilities.

it’s like giving your assistant a whole toolbox instead of just a notepad.

7/ Context engineering

Feeding the agent the right information, at the right time, in the right format.

it’s like prompt engineering but bigger combining instructions, real-time data, memory, and rules.

Better context = better decisions.

8/ Model Context Protocol (mcp)

A standard way for agents and other systems to talk to each other, even if they’re built differently.

Think of it like bluetooth but for ai systems sharing goals, data, and instructions.

9/ Langchain

An open-source toolkit that makes building llm-powered agents way easier.

You can chain tasks together, give agents memory, and plug in tools all without reinventing the wheel.

10/ Agentflow

A no-code, visual way to design custom agents.

You drag-and-drop workflows, set rules, and the agent runs it all.

Great for teams that want ai power without writing a single line of code.

I hope you've found this thread helpful.

Follow me @Yesterday_work_ for more.

Like/Repost the quote below if you can:

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling