Lilac CreativeMay 27, 2026

Agentic AI: What Nonprofits Need to Know 🤖

Agentic AI: What Nonprofits Need to Know 🤖

Lilac Creative

Learn the difference between standard AI chatbots and agentic AI, and understand why nonprofits should prioritize solid workflows and AI literacy before adopting autonomous agents.

Agentic AI: What Nonprofits Need to Know 🤖

Lilac Creative

Learn the difference between standard AI chatbots and agentic AI, and understand why nonprofits should prioritize solid workflows and AI literacy before adopting autonomous agents.

Chapters

00:00

Introduction to Agentic AI

An overview of Agentic AI and why it is relevant for nonprofit organizations.

00:43

Chatbots vs. AI Agents

01:24

The Agent Analogy

01:50

Real-World Applications

02:16

Understanding Potential Risks

02:56

Strategy for Nonprofits

Transcript

00:00

what is agentic AI and should non profits actually care?

00:04

So agentic AI, is the buzzword of 2026, and let me save you 47 LinkedIn carousels and just tell you exactly what it is and whether your nonprofit should care.

00:17

So, most explanations of this are useless.

00:21

They either talk down to you, or they assume that you've been following AI Twitter for, like, three years.

00:26

Neither of that helps if you're running a small team and you're trying to figure out what's worth your attention in the whole AI ecosystem.

00:34

So here is the version that I wish someone had given me early on, in plain English.

00:39

So the actual difference.

00:41

Here's the simplest way to think about it.

00:43

A, chatbot answers and an agent does.

00:48

So when you talk to ChatGPT or Claude, you are in the driver's seat.

00:52

You ask a question, you get an answer.

00:54

You ask another question, you get another answer.

00:56

Repeat.

00:57

Repeat.

00:58

Useful, but very, tedious.

01:00

Every step needs you, it needs you.

01:03

Prompting an agent is different.

01:05

You give it a goal, and it figures out the steps for you.

01:09

It plans.

01:10

It uses tools, searching the web, opening files, sending emails, running code.

01:16

It checks its own work, and then it tries again if something fails.

01:19

So you're not driving anymore.

01:21

You're more like a manager, overseeing the process, setting it up, and then overseeing the process.

01:26

So here's the analogy that helped me.

01:29

A chatbot is a really smart intern at the desk next to yours.

01:34

You ask questions, they answer, you take it from there.

01:37

An agent is an employee you sent across town with a credit card and your email password.

01:43

Both can be incredibly useful, and both can carry very different levels of risk.

01:48

Here's what that looks like in real.

01:51

The chatbot version.

01:53

Help me write a thank you email to a donor who gave $500.

01:56

An agent version.

01:58

Find every donor who gave $500 or more last year.

02:02

Write each of them a personalized thank you, and schedule the emails to send out Tuesday at 9am the second one is incredible, right?

02:10

If it works, it's also where things can go quietly, expensively wrong.

02:15

Wrong donor, wrong amount, wrong list.

02:19

And because the agent is taking multiple steps, one early mistake gets amplified all the way through.

02:26

So the real shift here isn't the technology, it's the unit of work.

02:32

We're moving from help me write this to go do this.

02:36

And that's genuinely a big change, especially for people who are just kind of getting into the AI space, right?

02:42

They're just trying to figure out which tool to use and what task to use it for.

02:46

So.

02:46

So the failure modes for both are bigger.

02:49

So an agent doesn't make one mistake.

02:51

It makes one mistake, and then it builds on it and then it sends it.

02:55

So it's really important when you are using an agent to test it before you just run with it.

03:03

So should your nonprofit care?

03:05

That's the big question, right?

03:06

Honestly, most small nonprofits do not need an agent yet.

03:11

You need solid workflows and AI literacy first.

03:13

You need an AI policy that your whole team has agreed on.

03:18

You need AI champions within different departments.

03:22

Those are things that come before AI agents.

03:25

Here's the test.

03:25

If you can't describe a process in writing, you can't hand it to an agent.

03:31

So the organizations actually benefiting from this stuff are the ones who have already mastered the boring parts.

03:36

Clear processes, clean data.

03:39

I know that's a tough one for nonprofits.

03:42

Good documentation, also tough for nonprofits.

03:45

So start with one question.

03:47

Where are we doing the same task by hand, over and over.

03:51

That is your candidate.

03:52

That's your test to start.

03:54

Not necessarily for an agent, but for any AI workflow at all.

03:59

This is the kind of thing that I keep breaking down inside the Lilac Collective.

04:02

So if you want one in your inbox each month, a little short video like this.

04:07

No hype, no fluff, just practical versions of AI.

04:11

Mission Shift is the newsletter that I also work on.

04:14

So click the link below and, subscribe.

04:17

See you in the next one.

Agentic AI: What Nonprofits Need to Know 🤖 — Tella