Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
Andy (00:00)
I mean we were talking in preparation for this episode.
one of the things that you said which resonates with me from when I
talk to clients is like really the agentic spectrum, like the promise
of agents and digital teammates is so palpable that when you talk
about it, business leaders suddenly are like, you mean and when you
talk about it and when they experience it and have a digital teammate
at work, it's like, holy cow. And so
We're still at the beginning of the agentic era.
Adam (00:28)
Right.
Andy (00:29)
the appeal is so great for business leaders, but I think you said
really this whole agentic spectrum comes down to a risk-reward
organizational calculation.
This is AI First with Adam and Andy, the show for leaders turning AI
from a tool of their testing into a workforce they run. I'm Andy Sack,
and alongside my co-host Adam Brottman, each episode we bring you
candid conversations with business leaders deploying real AI agents,
what's actually working, what's not, and how they're rebuilding their
companies around agents. No fluff, just real talk, actionable use
cases, and insights for you.
Greetings everyone. Today we are going to talk about the agentic
spectrum and try to give you a way to start to think about digital
teammates. And that's going to be informed by our
AI services practice because every day we're out there talking to
restaurants and retailers and other businesses, executives at all
those companies about how to use AI, but increasingly it's about how
do we use agents and digital teammates. So with that, I'm gonna share
a slide with all of you and ask my co-host Adam,
to talk about what is the agentic spectrum.
Yeah. So Adam, dive in here because like you're spending
Adam (01:46)
Yeah.
Andy (01:46)
a lot of time in our AI services. we recently wrote the book Agents
Inc., and we've been spending a lot of time thinking and talking about
digital teammates. let's define what the agentic spectrum is in its
three stages.
Adam (02:00)
Yeah. what I'll say is for the audience that can see the screen, I
think some people can see the screen and some people are just
listening to this. The screen shows a spectrum from you're not even
using AI all the way to like you have an actual digital teammate
that's powered by AI. And in the middle you've got call it, you know,
assisted intelligence and then agentic workflows as you get your way
to digital teammate. And
The thing that's kind of fun to think about now that was not really a
this time last year, last summer, we would not have been talking about
an agentic spectrum. We would have been talking about like, are you
using AI to augment what you do? Are you using AI to automate your
workflow? and that would be kind of it. And you know, are you using
AI? Are there applications where you're in integrating AI into your
applications? Now we like to talk about an agentic spectrum because.
It's turning out that your most basic tool that you're gonna use right
now, which is ChatGPT or Claude or Gemini, just right from the get-go,
they're so much more than a chat bot now. They are agents of a sort.
and so as you start to use AI right now, the more advanced you are,
and the more I'll even say willing to like take some risks around
like, are you willing to connect your email to it? Are you willing to
connect some sort of data source to it, even if it's only read-only,
which is safer.
Like the more you're sort of willing to enable your agentic platform
to do things for you, the more it's gonna feel like it's a teammate
than it is just a chat bot. And that's the biggest thing I want to
say, and turn it back over to you to kind of go back and forth with
me, is that we've said this a million times, we'll say it again. What
happened in the last six months is we went from a chatbot era to an
agent era, but it's not as simple as like.
I'm either using a chatbot or I'm using an agent. A lot of times your
Chat GPT, your Gemini, your Claude, they are both, depending on how
you're using them. And the more I'll call it data and power you give
to your AI system, the more it's going to be agentic for you and it's
going to be able to do more tasks for you, and maybe even multiple
tasks for you.
versus just give you a prompt and response. So Chatbot era, prompt and
response. Agentic era, give a task over to the system and treat it
like an agent and let it run and check its work and do multiple things
and then give you a deliverable. That's much different than a quick
prompt and response. And so really that's the biggest thing to keep in
mind and that's changed and the more you go down the spectrum, the
more
you're willing to sort of connect it to systems and give it memory and
schedule things to happen automatically, the more you're gonna make
your way from the left hand side of this chart that we have on our
screen here to the right hand side, all the way to eventually these
things become actual digital teammates, but that's the spectrum.
Andy (04:44)
And so I mean we were talking in preparation for this episode.
one of the things that you said which resonates with me from when I
talk to clients is like really the agentic spectrum, like the promise
of agents and digital teammates is so palpable that when you talk
about it, business leaders suddenly are like, you mean and when you
talk about it and when they experience it and have a digital teammate
at work, it's like, holy cow. And so
We're still at the beginning of the agentic era.
Adam (05:13)
Right.
Andy (05:14)
the appeal is so great for business leaders, but I think you said
really this whole agentic spectrum comes down to a risk-reward
organizational calculation. Can you talk more about that risk-reward
calculation
Adam (05:26)
Yeah.
Andy (05:27)
and why you say that?
Adam (05:28)
Yeah, like at a simple level, I'm gonna give an example. It's actually
a real world example of something for me today. So like if you want
ChatGPT or Claude to help you review a red line of a contract and your
lawyer's not available and you need to get the thing signed, and
you're just down to like, you know, four little changes that got
requested from the other side.
I hope the lawyers in the audience don't kill me. I hope our lawyer
doesn't kill me. But as an example of like, am I willing to stick the
contract into Chat GPT and then also maybe upload a download or
connect it to a folder on my system that has the email exchange. Maybe
even connect it to my email inbox and say, hey, look at my email
inbox, look at the back and forth from the last 24 hours, read the
original contract, and read the red line, and tell me.
How I should feel about these four changes today, because I'd like to
get it signed. I'm taking some risks there. I'm gonna be connecting
ChatGPT or Claude to my email, okay? Read only, but connecting it to
my email. I'm gonna be putting confidential information into a system.
but I'm gonna have an enterprise account and I know they're not
training in my data and I trust it. And it's gonna go and it's gonna
like kind of be my digital lawyer teammate. And it's gonna go read the
contracts and read the email and
How I should feel about these four changes today, because I'd like to
get it signed. I'm taking some risks there. I'm gonna be connecting
ChatGPT or Claude to my email, okay? Read only, but connecting it to
my email. I'm gonna be putting confidential information into a system.
but I'm gonna have an enterprise account and I know they're not
training in my data and I trust it. And it's gonna go and it's gonna
like kind of be my digital lawyer teammate. And it's gonna go read the
contracts and read the email and
And then go look at it and give me its thoughts on what I'm gonna do.
I'm taking some risk there. I'm not taking a lot of risk in my
opinion, because I'm on an enterprise system. I trust the
confidentiality. I know that I've set my connector to my email to
read-only. But there are people that are like, I'm not touching it,
I'm not connecting it to my email. Like, whoa, like what if that data
got out? What if it went rogued? How do I know it's read-only? How do
I know it's really confidential? By the way, these are real questions
we hear every day from our clients. But
I'm like, yeah, no, I've thought this through. I've got good
governance. Okay, that's an example of like risk reward. I don't need
to bother my attorney, it's not even available today. I'm just gonna
get this thing done. It's only four little points we can get the thing
signed. Then there's a more extreme example, which is like, you know,
am I willing to do some kind of like browser extension on my computer
with a co-work desktop app and let
Claude, take the wheel on a logged in version of some enterprise site
that I'm using and be like, hey, you know, you're logged in as me. Go
figure this stuff out and go even make some changes to this thing on
the site. Like, that's pretty risky because like it could make the
wrong changes. It's not read-only, but boy, that could save me days of
time. And if it's a recurring thing I need done, I might pick up 10,
20 hours a month of my time, depending on my job.
But I had to take a risk and I gotta know what I'm doing. that's a
more extreme example of like really letting an agent you know go to
work for me. But I'm getting more reward, but I'm also taking risk
there. And so I I feel like that yeah.
Andy (08:01)
Let me let me let me build on or add,
I'll add another example also from today, like real world, which was
just prior to this call, We have a financial analyst agent named
Barry. And we forwarded a small chains 15 location restaurant to
Barry.
And asked Barry to analyze the history. and Barry came back with a 10-
slide PowerPoint deck, and I forwarded it to Adam, and Barry did this
work, a digital teammate. I took the risk of forwarding the financial
history with, you know, under NDA to Barry.
And asked Barry to analyze the history. and Barry came back with a 10-
slide PowerPoint deck, and I forwarded it to Adam, and Barry did this
work, a digital teammate. I took the risk of forwarding the financial
history with, you know, under NDA to Barry.
Barry did this analysis and I was like, OMG. Like the analysis, Adam
looked at it, was like, holy cow, that chart looks great. I mean, what
a human employee would probably have spent the better part of at least
a week, probably two weeks, to get to this output. We got back in an
hour or less.
Because Barry, the digital teammate, the financial analyst, we trusted
it enough to give it this data and lickety split. Holy cow. And we
were able to make business decisions better, faster. That's the reward
as a result of the trust that we gave Barry. I mention all that
because going back to the agentic spectrum, on the left side, it's
very little use. The middle is automated workflow. And on the right,
Is digital teammates. In our book, Agents Inc., Adam and I lay that
out. You should go to agentsincbook.com and interact with Vera, who is
our co- agentic co-author. In our book, we talk about this move. This
move to digital teammates is going to happen fast. It's something that
you need to educate yourself on. We're talking to clients about it
every day, and it comes down to risk reward.
So that's my closing comment. Adam, you have anything you want to add
for our audience as I close up?
Adam (09:44)
I would just say that notice in every example we gave today, there was
no talk of not verifying the output, right? There was no talk of not
like having a human in the loop. And so even as we talked about risk,
the risk was more of like, did I set my system up securely? Did I
think about, you know, confidentiality, read-only, et cetera?
And so we're not advocating risk here. I just want to say that. But
I'm glad, Andy, that you brought this topic up, that, you know, you
gotta really start to lean into how these systems can use computers
and can connect to things if you're gonna wanna get more out of them.
And so I think the risk reward spectrum, the agentic spectrum, they go
together.
Andy (10:25)
Awesome. Yeah, I think in another mini episode we should probably have
an episode just on the risk side to actually define the risks and sort
of talk about that because
In our AI services work, that's a question that is always coming up.
In this episode, we really wanted to just give you a way to think
about the three-state agentic spectrum and recognize that it's a risk-
reward calculation, as well as give you a couple of real-world
examples. So thank you all for listening to AI First with Adam and
Andy. For more on how to become AI First, head to Forum3.com. You'll
find case studies, research briefings, executive summaries.
In our AI services work, that's a question that is always coming up.
In this episode, we really wanted to just give you a way to think
about the three-state agentic spectrum and recognize that it's a risk-
reward calculation, as well as give you a couple of real-world
examples. So thank you all for listening to AI First with Adam and
Andy. For more on how to become AI First, head to Forum3.com. You'll
find case studies, research briefings, executive summaries.
And our book, Agents Inc., How to Rebuild Your Company in the AI Age.
You'll also meet Vera, our agentic co-author, who's always up for a
conversation. If you want to connect with other leaders doing this
work, join the AI first community, our hub for turning AI hype into
action. We truly believe you can't overinvest in your AI learning.
Onward.