One year ago,
Salesforce launched
AgentForce to
redefine how we work
and interact with
our customers.
Since then, we've
been honing our
agentic tools
so our customers
can benefit from
our experience.
Kerry Chow, excited
to be joined by
my man, Chief
Digital Officer, Joe
Inzirillo. Joe,
thanks for being here.
In year one, what
were some of the
lessons we learned
about AgentForce?
Well, first off,
thanks for having me.
I think, you know,
AgentForce was, you know,
was born last year.
So it's hard to
imagine. It's only been
a year. But I think
one of the big things
that we learned coming
out of agent forces,
you know, we built a
lot of agents. A lot
of them were pretty
good. Some weren't as
good. But the focus
part of it, like
earning the right to go
fast, focusing on these
four hero agents that
we talk about. I'm
going to talk about my
keynote on Thursday.
I think it was really
critical is just
getting the reps down.
Like it's not like
developing other pieces
of software. These
agents are new. They're
a very different
type of thing. So you
really need to get the
reps in and that focus
on like getting some
critical use cases,
you know, getting them
to perform and then
just continuing to
ramp it up, throw more
traffic, add more
capabilities and just make it
better and better and
better. And that like
kind of continuous
improvement cycle, I
think, you know,
everybody's sort of new, but
like figuring out what
was the absolute right
way to do that with
an agent is a lot of
what we've done. And
we've seen the success
from it and we know
that it works. And then
we can start to scale
horizontally and take
on more and more use
cases. I like that. It's
like practice makes
perfect type of thing.
When you chat with
fellow execs and
leaders in the AI
space, especially
our customers
and our partners,
What are their
biggest concerns
and how are we
addressing them?
Yeah, well, I
think the biggest
concern that everybody
has right now is
really about data
and data privacy.
So these agents, you
know, they don't like
silos. They don't like
walls. So they want
to cross cut across
all your data sets and
things like that. And
obviously that's not
appropriate in all
cases. So I think the
biggest thing that
we're doing right now is
the trust and governance
layer is really
important. and being
able to actually make
sure that the agents
are getting access to
the right data for the
right customer, for
the right time, for
the right purpose, I
think is a huge industry
problem that I think
Salesforce, you know,
here, we have a great
solution to. So I
think that's probably
the biggest one. You
mentioned here at
Salesforce, how are our
teams using AI to enhance
productivity and how
has it changed the
day-to-day experience
for our teams? Well,
I think, you know,
every employee at this
point in time, we're up
to like 98% of employees
have used an agent
at some point and
something like high 80s
use an agent multiple
times a week of some
sort. So from the
employee experience
agent, you know, you go
to Slack, you're in the
flow of work, you're
actually talking to
an agent, you don't
even realize you're
talking to agent force.
So we have all sorts
of stuff there.
But I think if you
look at the things that
we're doing in the
areas that we really
focus like, you know,
the lead nurture
agent, I think is
just smoking hot.
And that's like a
great example of it's
really changing the
way the, especially the
small business side of
it is just approaching
things. You know,
we're, we're nurturing
leads in a fully
automated way for a lot
of people that we didn't
really talk to before.
So I think that's
massive. You know,
the engineering team's
got some great stuff
working with an
engineering agent that's
changing the way they
deploy code. So you have
all that sort of
stuff. And I think on a
day-to-day basis, the
way that we build the
things that we're
building is inherently
getting more and more
agentic. And And like
I said, whether that's
happening in the
flow of work at Slack
or whether it's built
in some of the other
tools like Tableau
Next, you can't turn
around and not touch an
agent at Salesforce
at this point in time.
It's wild to me. And
you kind of touched
on this before,
but like what did
Salesforce learn in
our adoption of agent
force that we've used
to help customers
do the same, like
the learnings that
we've applied now
to our customers?
Yeah, I think
enablement's a big thing,
right? Right. So, you
know, it's one of these
things that there's a
natural tension, which
is you want to do all
this iteration. You
want to make the agents
go as fast as possible
and learn and all
that. But at the same
time, human beings
still need to know what
agent to talk to to
go do something. So I
think, you know, not
having a million agents
is a good thing, trying
to get it to a central
point. But then also
really working with
the agents and the agent
design so that when
new features and things
like that pop up,
it can suggest things
that it can do. Because
if you're a salesperson
and, you know,
you've used it, you
may be very happy with
it, but it's like,
hey, I can provide an
account POV on that.
Would you like that, too?
Because how are you
going to get 75,000
people to understand
every new capability
you're going to
launch? And so I think
that's one of those
things that we're
getting more and more
sophisticated with.
And I'm hearing that,
you know, internally
that people are
really enjoying it.
But also, I think
our customers are
starting to see the
ability to do that, too,
as they go on their
journey and they
can continue to guide
and prompt their
customers using
AgentForce to outcomes
that they didn't even
know were possible.
Speaking of those
customer experience,
like how are we
helping them adopt,
integrate and now
use AgentForce?
Well, I think a big
part of what I do at
Customer Zero is
trying to pave the way.
So being, you know,
the early adopters of
our own technology,
getting a bunch of
learnings, working with
the product development
lifecycle to drill
those learnings
back into the product
is a big thing. And
then a lot of cases,
you know, we're
sprinting ahead of the
product and learning
things and then coming
back and informing
that perspective. So,
you know, I like to
say to our customers,
you know, I made a
thousand mistakes,
so you don't have to.
But it really is about
refinement and about
really, you know,
providing input in the
product development
lifecycle. But then the flip
side of that is also
coming out and things
like at Dreamforce,
doing a lot of panels,
doing a lot of
customer meetings, is
evangelizing what
we're doing and showing
them the path. It's
like every customer is
going to have a slightly
different journey,
but I can tell you
how we got started.
I can tell you what's
working for us.
And that practitioner
mindset really resonates
with our customers.
I love the fact
that you said we're
making the mistakes so
that you don't have to.
Really quickly, what's
next for Salesforce's
agentic transformation?
And it's wild to me
to even be saying
what's next. But here
we are at the pace
of which technology
moves. I got to know.
Yeah, I think there's
two things I'm really
excited about coming
up in the next year. I
think the notion of
agent orchestration,
where you have sort of
an Uber agent that's
actually able to do a
bunch of work with a
bunch of agents, which,
again, just narrows
the points that somebody
has to go to. At
some point, you'd love
just the Salesforce
agent and knew
everything about you and
could do everything
that every agent that
we've ever built could
do. So that's coming.
And no product
announcements, but that's
coming. So I'm excited
about that. And the
other thing that I'm
super excited about is
we're starting to get
more and more into
the process of sort
of redesigning the
organization around the
agents. So as the agent
impact goes up, we
can think about the
way that we organize
the humans differently.
And we're really
starting to see like that
becoming the agentic
enterprise is a
different enterprise. It's
not agents just bolted
onto the slide.
It's agents actually
helping us to find what
that organization is
going to look like. And
we're early innings,
but man, I think 27
or 2026, fiscal 27
is going to be really
exciting. And what
was the slogan you
were saying? Humans for
impact, AI for scale?
Yeah, I think that's
really what it comes
down to. You know, the
humans, the humans
are the ones that are
really driving the
impact. They're managing
these agents. They're
figuring out new
things. They're tuning.
But the agents are
scaling in ways that
are just unbelievable.
I mean, they never get
tired. They can work
24-7. They can speak
multiple languages.
You know, they're
instantly aware of
data changes,
things that are just
impossible for humans
to do. But the humans
and agents working
together, that's
the magic number.
Humans for impact,
agents for scale.
I love that. The
possibilities, limitless
like Bradley Cooper
back in the early
2000s. Absolutely.
Salesforce Chief
Digital Officer, Joe
Inzerillo, thanks so
much for joining us for
a step-by-step guide to
how your organization
can become an
agentic enterprise.
Scan the QR code
on your screen.