All right, we're
going to learn
how to put
Agentforce to work
for you from a
Salesforce leader who
knows the power
of Agentforce
From the keynote room,
straight to the desk,
Sanjana Parulekar,
VP, product marketing.
Thank you for being here.
I am so excited
to talk to you.
This technology, and
having been at Salesforce
for a really long time,
it feels so game-changing.
And not just for customers
and their customers,
but also for
their employees.
And so talking to the
community, a lot of times,
they just want to know
where do I get started
So let's start
there and talk
about what are the most
common use cases that you
And then, what are you
hearing from customers
Yeah, I mean, what I
love about Agentforce
is that there really is
a use case for everyone.
And we actually did
this exercise on my team
We're product
marketing team.
We have more work
to do than we have
So we thought, what is
the agent we would build?
And we said, we've
got to build an agent
to answer AE
questions on Slack.
Even if I hired
100 people,
we can never answer those.
And agents could
really do that best.
And so the way we
think about use cases
with customers is we
kind of model it out
based on what's
the role that you
In this case, it's the
go-to-market product
What's the knowledge
that it would need?
What's the data
it would need?
So in this case, it's
information from our Deal
Desk, our pricing
documentation, FAQs,
a lot of rich texts that
an expert go-to-market PMM
And then, what
are the actions we
want it to be able to do?
Because there are some
cases where you want it
You want it to
escalate to a human.
And then what is that
channel you deploy it on?
In this case, it's Slack.
So that's kind of the
framing of how we think
In terms of what
we're seeing,
I mean, we have
this going on
We've had this
at every event
where people are building
their own agents.
Service is a huge,
huge space for us.
As you know, just
general case deflection
and case management is
something that's really
Also in HR, HR type of
benefits, enrollment,
all that sort of stuff,
those types of workflows,
So for sales
teams, I mean,
SCRs, really, really
good SCRs and BDRs
And extending their
teams with Agentforce
is really the easiest
way to get started.
You mentioned
one that I really
want to dive a
little deeper
into because I think
it's pretty fascinating.
So answering AE questions.
So there's two nuances
to most AE questions,
their particular customer,
and then also that pricing
So when you think
about extending
some of those actions,
could an agent
ground itself in
that customer's data,
that specific account,
what they already
have, what their
terms are, what
their contract says,
interact with the AE
to make that
pricing proposal,
and then go through
pricing approvals?
Like, could you
essentially-- see,
that to me, is like the
possibilities are endless.
And that's the difference,
I think, with the actions,
being able to do
that pricing approval
because it meets
all the parameters
and it's aligned with what
that customer already has.
Think of how many days or
weeks that could shave off
And the way that I like
to make it really tangible
for folks is think
about the things
that us, as humans, we're
not that great at, right?
Are we really great at
recalling a 4,000 page
I don't know about
you, but that is not
And picking out
the exact phrase
that answers an
exact question?
No, but AI is really,
really good at that.
So let's put AI to
work at the things
that humans aren't great
at and don't want to do,
and let the humans
do the rest.
Yeah, because then
now, I spend the time
on the negotiation and
building that relationship
versus trying
to figure out
how to put this
proposal together
that is in line with
all those things.
So sales use cases are
really, really strong,
But service too, I
think, as a consumer,
I'm super excited
about these.
So let's talk about
the service use cases
and what are we hearing
from customers so far.
Because we saw some pretty
cool demos in the keynote.
Yeah, I mean, it's
all over the place.
And I think
the initial use
cases that you think
of are replacing
Like the Sophie demo that
we've shown at Dreamforce
that we just showed at
World Tour New York,
a very canonical use
case of making sure
those chatbots are
actually useful.
How many times have you
tried to use a chatbot
to solve a case,
and you just
give up before you
even get anything done?
And so what makes
agents really powerful
is that they're
fully conversational
and grounded in your data.
So that's one
type of use case.
I think what's also really
interesting, especially
for Salesforce customers
and going in this B2B
space that we're in,
is B2B service, right?
So think about if you've
ordered a washer/dryer
from a distributor
and they're
working with all
those vendors
to get services
done, and they're
So how do you empower
those businesses
with an agent to take
care of that end-to-end
Because to me,
the person with
the broken washer/dryer,
I don't care.
I just care that my
case is resolved.
So there's a lot of
possibilities and service
I think the field
service use cases
I love this one
because I have
had an experience
where I bought a brand
new refrigerator and I
had to defrost my turkey
before Christmas,
before Thanksgiving,
And I went out and opened
the freezer to remove it
and it already
did it itself.
And so the compressor
had gone out.
And the field
technician had
to come out
multiple times,
but it was simply because
of that connection
with the manufacturer
around that part.
And so being able to have
someone actually interact
with me as an agent
to do all those steps,
he might have only needed
to come out once, right?
And for me, that would
have been wonderful.
So I think about
how we take
those first initial steps
to take that action where
that agent could have
ordered the part, right?
Because it can learn
from in this instance,
in these cases, consumers
are typically having
So let's make sure those
things are on the track
I couldn't
imagine what that
would mean for not just
the technician cutting
from three
services to one,
but also that
manufacturer being
able to deliver
those parts, which
And what I love
about it too,
is that in these cases,
it's Agentforce solving
a data challenge,
and that technician
is going to come to your
house with that empathy,
that human-to-human,
like, I'm
so sorry that
this happened,
but I know exactly how
to solve this for you
because the HMS
helped me do so.
Yeah, and so you
brought up data.
And we've heard the
term don't DIY your AI.
And a lot of that
is rooted in data.
So let's talk a little
bit about Data Cloud
and the power
of Data Cloud,
and what role
that is playing
within the agentic
experience.
I mean, it's never
been more important.
I mean, it's the popular
adage of garbage in,
You need really
good quality data
I mean, our customers have
a ton of customer data
inside of CRM,
but to really see
that full picture of the
customer, it's everything.
It's their website
browsing behavior,
it's device data, it's
images, it's audio,
I mean, we're interacting
with businesses
in so many different
ways every day, right?
And so giving our
customers the ability
to use something like Data
Cloud that's lightweight,
that allows them to just
put their data to work,
I mean, we say that a
lot, but it's really true.
There's so much data
that's underutilized
And so with
Agentforce being
kind of that, like,
compelling new type
of application that
every consumer wants,
every business really has
more pressure than ever
to get their data
strategy in order.
And data strategy
sounds like, oh, my god.
I don't want to tackle
my data strategy.
And with Data
Cloud, they kind of
don't have to because we
have Data Cloud working
behind the scenes
with Agentforce
to pull all the right
data together, which is
And I liken this to when
you try to run a report,
in any system and you
run into a roadblock
because the data
point that you need
to connect to
make sense of that
or to take action on that
is sitting in another data
That has been a struggle
for organizations
And so the power
of Data Cloud
having it all
in one place,
now when you're running
a report, you just
select all of the
things, and being
able to then take
action on it.
And I think that those
agents having access
to that data, but I also
love the guardrails.
So just give us a
quick description
of the security and
why we're building it
in the way we're building
it and why it's important.
I mean, we love AI
and what it can do,
but as humans,
we love rules.
We love a little bit of
predictability, right?
I think everyone
can relate to that.
And so if I go back to my
example of a PMM member
There are going to be
instances when an AE
question comes up
that I don't actually
If it's maybe
for a customer
above a certain AOV band
or something like that,
I want to escalate that
immediately to Deal Desk
or to one of my
team members.
And so if you think about
the consumer service
You think about this
washer/dryer example.
There might be
an instance where
there is a natural
disaster, or god forbid,
I don't want an agent
handling any bit of that.
I want a live person
picking up the phone,
coming to it with empathy,
with that human-to-human
connection, and not
outsourcing that
So guardrails
really enable
you to put that into the
product do not do this,
do not ever escalate
this type of thing
or do escalate
this type of thing.
And it just gives a little
bit of predictability
Absolute legend, Sanjana.
Thank you so much
for joining us
today and sharing all of
those awesome insights.
