#8 - What is Agentic AI? From Annoying Assistants to Autonomous Action

We all know Siri kinda sucks—but what if your AI could actually get things done for you, without even asking? That’s the promise of agentic AI, and it’s not sci-fi anymore.

In this episode of Thriving in Ambiguity, host Steve Mancini breaks down how agentic AI takes things to the next level—planning, reasoning, and acting autonomously. From NVIDIA’s push to make it mainstream, to real-world examples in healthcare, customer service, and beyond—this is your guide to the next big leap in AI.

No fluff, just real talk on how AI is about to stop answering questions—and start solving problems.

Episode Transcript

(00:06):
Hey everyone.
Welcome back to Thriving In Ambiguity,the podcast where we break down
the complexities of technology,ideally without losing our minds.
I'm your host, SteveMancini, and at this point.
Most of us know about generative ai.
Think open ai, Google, Gemini,and clawed from anthropic.

(00:27):
These fantastic models can answeryour questions, do your homework,
or even generate cool new content.
But what if your AI couldanticipate your needs?
Plan your next steps andtake action on your behalf.
No finger lifting required.
That's the promise of a genetic ai.

(00:47):
So let's dig into the next frontierof artificial intelligence.
Now, let's be real.
We've all wondered, why does Siri suck?
Have you ever asked a question and itgives you a completely useless response,
making you wonder, why did I evenask that question in the first place?

(01:08):
Well, in the not so distantfuture, that response and
experience may actually be useful.
Imagine having an AI assistant thatdoes more than tell you the weather.
It actually gets stuff done.
Agentic AI uses sophisticated reasoningand iterative planning to tackle complex,

(01:29):
multi-step problems autonomously.
It's not just a chatbot thatspits out quick responses.
It pulls in data from multiple sources.
It formulates a plan and it executeson it, and it even learns from the
results so it can do better next time.
This is honestly what Ihad hoped Siri would be.
I mean, imagine if Siri hit you up andsaid, Hey, Steve, today I unsubscribed you

(01:55):
from all of your junk mail in your inbox.
Here's a summary that should helpclean things up a bit for you.
I mean, come on.
That would be awesome.
It's just, unfortunately that's not whatit is, but hopefully it'll get there.
I'm sure it will.
Now, if you somehow missed it.

(02:17):
Behind the scenes, Nvidia is thecompany that makes the hardware,
that makes advanced AI possible,and because of that, they are all
in on accelerating AI adoption.
And then they've createdblueprints for Agentic ai.
And you say, well, why?
Why are they so dialed in on this?

(02:38):
Well, because the more companies thatroll out AI solutions, the more computing
power and thus hardware they'll need.
By providing the sample apps, thereference code and best practices,
NVIDIA lowers the barriers to AIadoption, which in turn drives demand
for their high performance GPUs.

(02:59):
So basically NVIDIA's like.
Let me make this simple people, andoh, by the way, give me your money
and let's keep this train rolling.
So at this point, you know, you maybe thinking, okay, well, what does
agentic AI look like in the real world?
Take a customer servicedepartment, for example.

(03:20):
Traditional AI can answer basicfacts, but Agentic AI goes deeper.
It sees a customer's overdue balance.
It calculates which accounts can coverit, and all the customer has to do is
confirm and it'll take care of the rest.
Maybe a little bit betterscenario be in healthcare.
It may comb through enormouspatient data sets, spot anomalies,

(03:45):
and tee up a doctor's next stepsbefore they're even requested.
Might sound something like this.
Hey Doc, have you thought of this?
Oh.
And here's the best options for treatment.
Let's be honest, there's only somuch even the best doctors can know,
but what if their AI assistant hadaccess to all of the case data?

(04:07):
Every medical journal, you getthe point that potential for
lifesaving treatments explodes.
See AG Agentic AI thrives ondata both internal and external.
NVIDIA's end-to-end platform, includingtheir NEMO microservices helps developers
integrate that data, train AI models,and deploy them responsibly with

(04:31):
guardrails for safety and compliance.
In other words.
NVIDIA provides the software ecosystem socompanies don't have to reinvent the wheel
every time they want to implement ai.
The bottom line Agentic AI is setto transform industries, whether
it's automating code writing tasks,or coordinating supply chains.

(04:52):
By offering these blueprints, NVIDIA makesit easier for businesses to jump on board.
And as AI becomes a go-to resourcefor daily operations, the need
for robust hardware grows andso does their stock prices.
So ready or not Agentic, AI ishere and it's unlocking a whole new
level of productivity and insight.

(05:14):
The real question isn't if you'll use it.
But when, or put another way, it's nolonger a matter of if Siri will become
actually helpful, but when it stopsbeing annoying and starts acting like the
agent of assistance we've always wanted.
Thanks for joining Thriving In Ambiguity.
I hope you've enjoyed this episodeand please like and subscribe

(05:35):
so that you can catch the nextepisode on your favorite platform.
Thanks, and I'll talk to you soon.