The two most
powerful words
in the world that we
know are, what if?
Every great thing that
I've ever been a part of
began with
those two words.
And today, thousands of
organizations are saying,
With the power
to act on those
What if every
child had a tutor?
What if every health
care visit felt personal?
What if every family
had their own financial
What if your business
never had to sleep?
What if your company could
solve problems before you
What if you could
empower every employee
with a limitless
digital workforce,
and that listens,
learns and acts,
Well, this isn't some
far-fetched science
It's happening right now
with AI from Salesforce.
Agentforce turns your
vision into reality.
A future where every
student, patient, client,
citizen, and customer
experiences the best
Well, now they're free
to focus on the work that
And all you have to do
is bring your dreams.
What if every company was
an Agentforce company?
Please welcome
co-founder, of Salesforce,
and Chief Technology
Officer, Slack.
Welcome, everyone
here in the room.
I think it's a full house.
Look, we got
people back there.
And thank you to
everyone online.
Welcome to World
Tour, or Agentforce,
Agentforce World Tour, DC.
I'm so excited
to be back here.
I've been dedicated
to the federal market,
especially for Salesforce,
for well over a decade.
And Dave Ray,
if you're here,
correct me if it's
even longer than that.
But we've been working
together here in DC
to really drive
a lot of change
here, and help make the
federal government even
more efficient and better
to serve our citizens.
But before we get
the show started,
we're going to
do what we always
do, to thank each
and every one of you,
each and every
one of you who
are commercial customers,
our federal customers,
our SI partners,
nonprofits.
If there are
nonprofits in the room,
give a shout-out
to all nonprofits.
Thank you for being here.
Thank you to all
the stakeholders
who are part of
our community.
For 26 years together,
we have built something
And we couldn't have
done it without you.
But as you know,
26 years ago
when I met Marc Benioff,
we started the company
In 1999, you may
not remember,
but people didn't trust
their credit card number
They were like,
I don't know
if I'm going to
put my credit card
Much less my
customer list.
Trust was
fundamental in 1999.
It's still
fundamental today.
Today, it's more about AI.
We're not going to use
your data in our products.
We are not going to use
your data to build models,
to give back to
other customers.
That's everything
about trust.
But what's more important
than innovation,
With the technology
that you have today,
we want to make
you successful.
And if you're
not successful,
I hope you'll find a
Salesforce employee,
find me after the show,
tell us what's wrong.
Tell us what we can do to
make you more successful.
Everything about
today, you're
going to see some
incredible innovation.
And of course, we
believe a world
based upon equality
and sustainability.
And with those values,
we have created a company
that has driven
a lot of change,
that we believe that
business is the greatest
Well, when I met
Marc, he said,
we're going to start
a company with the 111
And I was like, Marc,
what is the 111 model?
And he said, well, let
me tell you, Parker.
1% of our equity,
which was worth nothing
at the time, we're going
to put into a 501(c)(3).
Almost $800 million
has been given away
by Salesforce in all
time giving to do good.
But it's not
just about money.
It's fine to
give money away.
That is a very good thing.
But it's also--
and this is
what we're going to
talk about today.
It's about human capital.
And we put the
human capital
of Salesforce employees,
they can get nine days off
I was just talking
to Marc Benioff.
He's off there somewhere.
In Japan, in
San Francisco,
wherever we
have a presence,
our employees
are giving back.
9.6 million hours
giving back.
And if you're a
nonprofit in the room,
60,000, or more than
60,000 companies are using
our product for
free, or near free.
You probably would
like it entirely free.
It's highly,
highly discounted.
Any money we make
on that goes back
We're now using
that capital.
We're only using it
to keep that going.
And also, today,
I'm super excited.
We are making $1
million donation
to American Forests in the
spirit of reforestation.
Because we need
more trees.
We're doing well
and doing good.
We can do both
at the same time.
When I met Marc,
he said, we
are not going to
create this nonprofit
and give back once
we're successful,
we're going to do it
every step of the way.
Those values are baked in.
And I believe it makes
us a better company.
40.9 billion in projected
revenue this year.
But even better, world's
most innovative companies.
Top 100 companies that
care in philanthropy.
World's most
ethical companies.
Those are all
shared values.
That's doing well and
doing good, all together.
And as we're doing
well and doing good,
it's about serving
all stakeholders.
And here in DC,
I'm especially
reminded of the
stakeholders of state
I'm right now dedicating
myself in San Francisco
to do a lot of work
locally in the San
Francisco market
with our new mayor
to improve the state
of San Francisco.
And we're going to do
that with Agentforce,
with Data Cloud, with
everything we're talking
about today, but
also 50 states,
not just San Francisco,
100 federal agencies,
60,000 nonprofits that
we just talked about.
And there's a lot of
those organizations.
You're going to
hear a lot more
about that
organization later
with a great video,
a great story.
But I want to bring it
home here into the room.
And I want to tell
you a little story
about the United
States Army.
I'd like to welcome
Bob O'Brien.
I keep thinking there's
some slide coming up, Bob.
We're supposed
to stand here,
so that people can see
our beautiful logo,
and we can look
at the cameras.
Thank you for your
service, by the way.
So, Bob, I don't think
everyone knows who you are
and what you're doing, but
it's around human capital
Maybe tell everyone
what mission
We got a whole army
of people to help you.
I'm a 32-year veteran of
the United States Army.
And I've been hired
back after retiring
last summer to continue
leading a program called
the US Army Command
Assessment Program.
We're the largest leader
assessment program
And every year, we bring
in over 2,000 candidates
to assess them over
a four-day period,
and make a selection for
putting them in command
and key leader positions.
And we were talking a
little bit backstage
A lot of data points
that you're tracking.
It's not just what other
leaders are reporting,
but there's some
psychometric data.
When you think
about data--
and also, like we're
going to talk about AI,
the potential maybe of AI,
how that might help you.
Maybe talk a little
bit about where
you see this going,
and any advice you
So across these
2000 candidates,
we've been assessing them
for over five years now.
So we've got 9,000
candidates that have come
through our
assessment program.
We gather over 3,000 data
points on each of these
candidates using a
series of psychometric
assessments, personality
assessments intelligence
And then we do
on-site work samples.
And we expand
what has been,
prior to five years ago,
a selection process that
considered every
individual for about
And we've expanded that
to 17.5 contact hours
So we now have over
27 million data
points on key
leader behaviors.
And we're gathering that
in a clean manner, which
is a great first
step for us.
We've taken what used to
be a very dirty data pool.
And we've cleaned
that data pool
And we're making
selections
on whether somebody
should be certified,
qualified to do a job,
or not yet certified,
disqualified for a period
of time, to do a job.
And as we do that,
about 13% of our people
And we have a large
leader development program
that's tailored
to the individual.
We use data to help
develop the leaders,
and prepare them to come
back and be successful.
And we've got a
number of cases
where leaders have come
and been unsuccessful
one time, and come back,
and been very successful.
Those leaders that
get certified.
Once they do, we're
able to take the data,
and not only use that for
the selection process,
but we can use that data
to deliberately build
teams of people
using the data.
And we just finished
our slating process,
and briefed the chief
of staff of the Army
And we've had some cases
where we have deliberately
recommended against
partnership teams coming
Because we had
leaders who had
the exact same personality
and behavior and makeup.
And we broke them
apart, even though they
were certified, because
the same stimulus
would cause them to
act the same way.
Someone once told me,
we did an analysis,
that I'm the complete
opposite of my wife,
Thank you for that story.
Thank you for
your partnership.
We'd love to
help you more.
Thank you for being here.
Well, I love
that story, Bob.
Because the one
point you made
of not everyone makes
it the first time,
but you're
investing in them,
and you're
bringing them back.
And it's about
that human capital.
It's not about getting
rid of employees.
It's not about
rejecting them
if they don't
pass some test.
It's about
investing in them.
And this is a
great example
that customer success is
driven by human capital.
It's driven by
human beings.
That if you invest
in those humans,
the frontline workers or
the engaged employees,
you will have more
successful customer
interactions, or
citizen interactions.
Well, that's great, but
everyone's overwhelmed.
Here in DC, we're
having a lot
of talk around efficiency,
which is wonderful.
And that means
potentially less people
And the people that
there are worried.
They're like, whoa,
how are we going
Do we have all the
talent we need?
There's a fixed capacity.
I wish I could hire
as many engineers
as I could to give you all
of this innovation even
These are all
issues you're all
seeing in your
organizations,
and in your
government agencies.
People don't care that
your talent is not
That you have people
that are burned out,
They expect you to
know who they are.
They expect zero hold
time when they say,
I'm frustrated
with some problem.
And they want to, when
you answer the phone,
or you answer their
text or their chat,
or however you're
interacting with them,
you better know
them, and you better
I hate when I call some
customer service line,
and they're like, the
first answer is like, no,
Please put me with someone
who actually can help me.
We've all had that,
but we don't like that.
So workforces
are overwhelmed.
That is the problem
that we all see today.
But what if workforce
had no limits?
What if there
were no limits?
And it's not
about no limits in
like, we're going to
all hire and create
all these incredible
employees.
It's about humans and
AI working together.
And that's what we are
doing with Agentforce.
That's what we
have been building.
You all know all
of our clouds.
We started with
Salesforce automation.
We bought marketing
automation.
But we've rebuilt all
those applications.
It's now one deeply
unified platform
of Customer 360
applications.
Thankfully, we built that
integrated Data Cloud.
Because data is the
lifeblood of AI.
It's the lifeblood of
personalized customer
It all works on that
unified Salesforce
You don't have to learn
multiple platforms.
You don't have to be
an expert programmer.
You can use that platform,
powered by Data Cloud,
through all those
touch points,
to give agentic
experiences, and basically
digital employees
to your workforce,
so that you can
scale infinitely,
that it will be
trusted, Easy to deploy.
The humans are right
there working with the AI.
We're not getting
rid of the humans.
And you can solve
that zero hold time.
And it's also
completely open.
We're not this
closed platform.
So we know that
we have to fit
Agentforce is powering
this digital labor
You think about
hiring employees.
I think about Bob, making
those employees better
leaders, better
at the skills
that they need
to be successful.
But you're also
going to start
hiring digital employees,
which is this weird idea.
You may actually fire
digital employees, which
Maybe there's an agent
not doing such a good job.
But you're going
to start thinking
about the legal agents,
the financial agents,
the developer agents, the
customer service agents.
These are experts
in your organization
that you're
going to train.
And you're going to know
them better than I will.
You know your business
better than we do.
So those humans
and all the IP,
you're going to imbue that
into this agentic platform
to create these
digital employees,
Salesforce is
doing just that.
We've been going
as fast as we can.
I think you saw the
pace of innovation
from Salesforce, from
everyone in the market.
We're going as
fast as we can.
It's a race right now
to take the power of--
we're all shocked by
the power of these LLMs.
The power of AI has really
hit that tipping point
We have to go as
fast as possible.
But we're all a
little scared.
Salesforce was
a little scared.
We're like, I
don't know if we're
quite ready to put a
Agentforce on our website
to help all our customers
because, hey, there
might be a few
problems with it.
Marc said, nope, you're
doing it right now.
And it didn't work
perfectly at first.
We put it out there,
and we made it better,
Every time there's
an interaction,
because we have
this platform,
We know if
there's a problem,
if a customer didn't
have the right answer,
We'll talk about
that later.
It's not just
a little LLM.
And with that,
we've driven 30,000
Our agents are talking
to our customers,
and they are solving
those problems.
85% of the time,
customers, they're happy.
15% of the time, those
are more complex problems.
And so we had
a human come in
to have that higher value
interaction to really help
And it was humans
and AI together,
solving and
scaling Salesforce,
because we can't hire
enough people to support
We're also doing
that in sales.
25,000 sellers have a
dedicated sales coach.
It's like if you could
hire the best sales coach,
and put them side by side
every employee that's
selling, wouldn't
you want to do that?
But you can't hire
enough people to do that.
But now you can
put the sales coach
And the best
thing about it
is, the salespeople
aren't embarrassed to ask
questions of
that sales coach,
because it's not
a human being.
So it's this weird
thing that they're
3.5 thousand exec briefs
generated by Agentforce.
So that's what we're
doing at Salesforce.
In our customer base,
RBC Wealth Management
has an issue where
they want to make sure
that their wealth
advisors are
serving their clients
the best way possible.
And how do they
scale that?
Well, they scale that
with an agent side
by side with that
wealth manager,
because it's a highly
personalized interaction,
even better and more
powerful with the power
of Data Cloud,
powering Agentforce
for those wealth advisors.
How many of even know
about ReMarkable?
It's these
amazing tablets.
It's basically like paper,
except it's not paper.
OK, I need a little help.
How do I-- there's
gestures in it.
I need to have the text
be recognized online,
and my writing
turned into text,
because I want to save it.
Or a problem
with my account.
They couldn't scale
the support enough,
fast enough, so they
put agent forests right
in the middle of
ReMarkable support.
And now it's deflecting
35,000 inquiries,
solving the
problems faster.
That's what we want,
faster and better,
so the customers
are happier.
Well, we're
going to show you
how you can do that
in your companies,
how you can do it
as a nonprofit,
as a commercial
enterprise,
as a government
contractor,
We're going to
show you how
to do that, how to do that
across the Customer 360,
How to make it more
personalized with Data
Because data is
the lifeblood,
as I've said, of this
personalized interactions
And most importantly,
this future of AI.
How are you going
to go quickly
We're going to do
that with Agentforce.
And to tell you
more about that,
and show you some
incredible demos,
I want to invite
Avanthika Ramesh.
I just mispronounced
your name.
Avanthika Ramesh,
senior director
So glad to be here
right at the cusp
of cherry blossom season.
And if you think these
cherry blossoms are
blooming fast,
well, Agentforce
Agentforce is a
massive transformation,
not just for us
at Salesforce,
but also for every single
one of your businesses.
This transformation
we're seeing today,
reminds me of a revolution
that we saw 50 years ago,
with the advent
of the CPU,
the central
processing unit.
For the first time,
we had a system
that could process inputs
and perform computations.
But what good was
a CPU on its own?
It needed an
operating system.
It needed all
this scaffolding
around the CPU,
which is what
made it usable
as a computer,
and led to its
mass adoption.
And now, here
we are today,
with a new revolution
of large language
For the first
time, we have
systems that can
understand reason,
and respond with
natural human language.
But just like a CPU, how
good is an LLM on its own?
We need to build on top
of these large language
models in order to unlock
their full potential.
And we need to feed
your organizational data
in real-time to
these models.
These LLMs are trained
on a finite set of data
So the magic is
in the prompts.
The prompts that you
send to these models
They can be grounded
in your structured and
unstructured data,
your metadata,
even the underlying
semantics of your data.
And with good prompts, you
get relevant responses.
But you also want to
surface these responses
And that's where the
application layer
We want to make sure that
your AI is context-aware
and well integrated
into your workflows
But we know that AI can go
beyond simply generating
It can even take actions.
And that's where AI agents
come into the picture.
Agents can reason
on top of your data,
on top of your
applications,
in order to take actions
across your business
Now, as you can
see here, this
is a lot to bring
together in order
That's why we've done a
lot of the heavy lifting
for you, so you don't have
to make those investments,
and brought all these
layers and services
into one deeply trusted
and unified platform
to make AI usable
for your enterprise.
We've also designed
this platform
so that it's as
open as possible,
allowing you to connect
to your own large language
models, allowing
you to connect
into your own systems
and APIs, even
access your data all
around the world.
This is that
single trinity
of data, apps and agents.
And we call this
platform Agentforce,
the digital
laborer platform.
We built this
agentic layer
on top of your data
and your applications
to help you build
these agents quickly,
and realize value
16 times faster,
with 75% more accurate
and relevant responses.
Now, we've designed
agent force
to be well integrated
into your business.
Look at all those zero
copy data partnerships.
This allows you to connect
and access your data,
regardless of where
it is in the world.
Then we have these
AI partnerships.
We allow you to
access a whole host
of large language
models across vendors,
or even connect
your own models.
And with our new strategic
partnership with Google,
we're bringing Gemini,
Vertex, and Google
Cloud straight
into agent force.
Now, something new you
might have noticed here
We have over 200
partners who are actively
building agent
templates and pre-built
actions for you, to help
you deploy these agents
And what you're seeing
here is [AUDIO OUT]
For AI applications
in the world.
Now, speaking of
actions, Agentforce
can take actions, both
within Salesforce,
and outside of Salesforce.
For example, it
can qualify leads,
it can schedule
appointments, or even
update inventory
on your behalf.
On top of that,
it can even
take action inside
of Salesforce.
For example, it can
process a refund in SAP,
or onboard a new
employee in Workday.
So now that you've
learned about the platform
and the ecosystem,
what about
Well, first, it starts
with some sort of trigger.
This trigger
could either be
some sort of anomaly
in your data,
a change in
customer behavior,
Then we have the
reasoning engine
that powers Agentforce,
that's able to assess
the context, gather
more data on demand,
and generate a
plan of action.
Now, this plan could be
as simple as generating
a single-shot
response, or as
complex as
orchestrating actions
And finally, we
have our outcomes.
Our agents able to
orchestrate these actions
in order to complete an
entire business workflow,
Now, as we look at
Agentforce, and all
the capabilities, one
thing I love to see
is how this comes to life.
My team has been
hard at work building
And it's gratifying
when our customers can
put these innovations
to use, and innovate
I've had the
pleasure of working
with wonderful customers,
like Fisher and Paykel,
Fisher and Paykel is known
for their high quality
products and
premium services
Now, as their
business grew,
it was equally important
for their customer service
to scale, and be just as
premium as their products.
And that's why they
turned to Agentforce,
so they could deliver
these exceptional customer
service experiences
at scale.
Now, before we get into
what Fisher and Paykel has
done, I have a quick
side story to share.
Last month, I was going
through a renovation
And for anyone going
through remodeling,
you want to make sure
you measure twice
and order once, not
the other way around,
Because of course, I ended
up with the wrong size
Now, the movers
were in the room.
I was trying to get a
hold of my refrigerator
company, and literally get
on call with the agent.
Now, I'm sure
you all have been
So let me
actually show you
what that experience
was like for me.
So I'm going to go ahead
and call customer service.
And thank you for
calling customer service.
If you know the
extension of the party
you are trying
to reach, you
To hear our store
hours, please press 1.
To start a return,
please press 2.
I'm sure you all
can relate, right?
And especially when
I'm in panic mode,
just trying to get
a hold of an agent,
this kind of experience
is just not helpful.
In fact, there is a term
in the service world
for this type
of experience,
and it's called
deflection,
which is essentially
deflecting
the human inquiry away
from a human agent.
But of course,
the customer
doesn't get the
help that they need.
Well, the good news is,
customer service doesn't
Because we have
customers and businesses,
like Fisher and Paykel,
who are at the cutting
edge and ready to
transform this experience
So now let's actually see
what my experience could
have been like
if I had instead
bought my refrigerator
from Fisher and Paykel.
I'm Will, your Fisher
and Paykel AI assistant.
How can I help you today?
So I had accidentally
ordered the wrong size
I'm afraid it's
a little small.
So is there any way I
can get a replacement?
That sounds frustrating,
but I'm here to help.
Is this about the 17.1
cubic foot Series 7
refrigerator you
purchased on March 3rd?
I have a couple
larger capacity models
in mind that offer
the same benefits
Are you able to take
a picture of the space
so I can recommend
the best option?
Yeah, I have a
picture on-hand.
How do I get
it over to you?
I've just sent you
a text message where
Let me go ahead and
select my picture,
and just send
it over to you.
Based on the dimensions
of the space,
I recommend the 20.1f
cubic foot Series 7
It offers plenty
of flexible space
for cold and frozen
food storage.
Would you like to
go with this option?
How soon can you get
it delivered though?
The soonest delivery
appointment available is
March 21 from 12:00
PM to 6:00 PM.
Yes, Friday works for me,
but I do have a dinner
Is there any way that you
could get it delivered
between 12:00 to
3:00 PM instead?
Unfortunately, I'm
not able to reschedule
If you would like,
I can connect you
with a representative to
explore other options.
I'll get back
to you, Will.
Right off the bat,
I'm sure we can all
agree that experience
was completely
First of all, Will
knew who I was.
He was able to
look up my order
and process a
replacement for me.
But of course, it
could have been better,
because Will didn't have
the ability to reschedule
Well, we don't
have to stress.
Will didn't have a
capability agent.
So what we're going
to do together
is actually go
into Agentforce,
and see if we can give
Will this new capability
So let's do this together.
We're going to go
into Agent Builder.
This is where all
the magic happens.
What you see here
is our agent, Will.
We start by seeing
a description
He's an AI-powered agent
for Fischer and Paykel.
And we also see his role.
And we can also see
all the channels
he's activated on,
like Apple messages,
email, Facebook
Messenger, even Voice.
Let's get into the topics.
These topics are
skills, or jobs
to be done that
Will is capable of.
As you can see here, he
can answer customer FAQs,
he can answer
general questions
and manage
orders, and even
answer to product
inquiries.
But as you can
see here, he
doesn't have the ability
to reschedule orders
So we're going to
go ahead and design
So let's go ahead
and hit New,
Now, all I need
to do here is
describe what
new ability I
want to give Will
in natural language.
Now, I'm not going
to do the typing
in the audience, so I
have it pasted right here.
We want Will to manage
product deliveries
Now, once I hit
Next, you'll
notice that
Agentforce does a lot
It's able to craft
that into a topic,
describe it so that the
reasoning engine knows
when to invoke this
topic, and it's also
able to describe
the scope.
These are the guardrails
of what it should
And finally, we have
instructions, more
specifics around the
brand voice and tone
So now that we have all
these details here--
again, all described
in natural language--
we're going to go
ahead and start
prescribing actions
within this topic.
Now, what you'll see here,
it's recommended some
out-of-the-box actions
that are most suitable
But on top of
that, it's also
looked at my
existing investments
in the platform, my flows,
my prompt templates,
And it's able
to craft them
into actions for my
agent, as we can see here.
So I went ahead and
selected those for me.
Let's go ahead and
create this new topic.
So I see that new topic
added right there.
Let's quickly review
all the details.
The configuration
looks good.
So remember I was asking
Will to reschedule
So let's actually
see if Will
is capable of doing that.
So I'm going to
type in my name,
and see if Will
can deliver my new
refrigerator between
12:00 to 3:00 PM.
It's about the
20.1 refrigerator.
And as Will is thinking
through my question, what
you see here in the middle
is our reasoning engine
You get full transparency
into how your agent is
reasoning, how
it's classifying
these utterances into
the right topics,
and orchestrating the
right set of actions.
Will is able to
reschedule my delivery.
And on top of
that, I can see
how he went through
and navigated
all these actions
in order to come up
And I can even provide
feedback to my agent.
So I have data to continue
refining and improving it
So now that we tested
this within Agent Builder,
why don't we give will
a call one more time,
and actually prove
it out to you?
I'm Will, your Fisher
and Paykel AI assistant.
How can I help you today?
I wanted to see if
it is still possible
to reschedule my delivery
to Friday between 12:00
Is this about the 20.1ft
cubic foot Series 7
I was able to move
a few things around,
and secured a slot for
your fridge delivery.
Derek J will be there
between 12:00 PM and 3:00
Is there anything else
I can help you with?
Within a few minutes,
it was that simple.
I wrote no lines
of code, and I
was able to give
our agent, Will,
this new capability
to manage
orders and reschedule
rescheduled deliveries.
And if I can do this
without writing any code,
trust me, you all can too.
So I hope you can join my
team out on the campground
after this keynote to
get hands on, and see
this magic for yourselves.
Thank you all so much
for following along,
and I'm going to hand
it back to Parker.
And I love the
demonstration.
I love the demonstration
with voice, too.
It's really,
really powerful.
So you just saw
a great story
of how to build
with agent force.
Unfortunately, Avanthika
didn't have enough time
to demonstrate everything
we have in the platform.
We actually just
had our conference,
as I mentioned earlier,
a developer conference.
And so you saw how is
the reasoning happening.
We saw in that
demonstration.
But there is so much more.
One of the things that I
love that we just came out
with is the
Developer Edition.
So all of you can
now actually do
Agentforce and Data
Cloud for free,
and learn it in
Developer Edition.
Also, how am I going
to develop this,
and how am I
going to debug it?
How am I going to test it?
We have a testing center.
It's a full,
full platform.
That's what we mean by
it's not just the LLM,
I'm sure you can see more
of it on the expo later.
But we're not going
to let you go yet,
because we want
to talk about
how does data power AI,
and how are you going
to unlock all
of that trapped
And to really tell you
all about that, please
bring up MaryAnn Patel,
SVP of product management.
Leading product for Sales
Cloud, talking to you
is the favorite
part of my job.
And a question that
I often hear is,
hey, do I have to train an
LLM on my organization's
And honestly, thankfully,
the answer is no.
This is really hard
for most of us.
And frankly, a
lot of people
don't have the
resources to do that.
But that's why I'm
thankful for Parker.
Because we have been
making the hard things
easy for the
last 26 years,
and we continue to do
that with the prompt.
The prompt, like
Avanthika just showed you,
is simple, conversational
instructions
And that is how you get
value out of the LLM.
Even better, we
have added retrieval
augmented
generation, which
is a fancy way of
saying we retrieve
the information,
structured and
unstructured, from all
of your data sources.
We augment the prompt
with your data.
We bring your
data to the LLM.
And that's how you
unlock the value.
And then we generate the
trusted, accurate response
that you need to deliver
that amazing personalized
This is the power
of Data Cloud.
But I know you're
all thinking, ooh,
there's probably some
data that AI needs that I
don't have access to yet.
And that is what we see
with a lot of customers.
A lot of your
data is siloed.
Locked data is
unfortunately
But that is why we
have built Data Cloud.
We are incredibly
excited about this,
because this is your
hyperscale engine
We allow you to bring
all of your data
lakes, your ERP,
spreadsheets,
You can pull it all
into Data Cloud.
We have hundreds
of connectors
to make it easy for you,
for you to connect that
in, and then you're
bringing that
Data Cloud is
the foundation
And it's incredibly
important.
But I know the
one thing that
is on everyone's mind in
this room, which is, hey,
Just like Parker said,
your data is my data.
I want to make sure
that I protect it.
And that is why we've
built robust governance
and data security
right into Data Cloud.
So you can feel
safe that your data
And the way
that we do this
is with Agentforce
and Atlas.
So if you think about
that reasoning engine
that Avanthika was
showing you in the middle.
This is what
that looks like.
It looks pretty
complicated,
maybe intentionally
complicated,
but I'm going to
make it easy for you.
In the upper right,
that's Data Cloud.
That's where
you bring in all
of your structured,
unstructured data.
If we start on the left,
you ask a question,
you gave it some
instructions,
and then we're pulling
it into the planner.
So let's think
what a human does.
When you ask me a
question, first I think,
what's the contextually
relevant data
that I need to
answer the question?
And then I create
a plan, and I
think about what are the
steps that I need to do.
And I'll go through
that a couple of times,
and make sure that
I'm iterating through,
so that we get
the right outcome.
That's exactly what
Agentforce is doing.
It's pulling all of
that into the planner.
And it is unlocking
the value in your data
Vivint is an agent
force company.
They are a home security
and automation company
that provides 24/7
customer service
and proactive monitoring
and installation.
They deeply care about
scaling their business,
And that is why
they use Agentforce.
The best thing
is, they are not
They have invested
in their processes,
their business,
their data.
It's all inside
of Salesforce.
For them to
use AI, it just
So now to show
you a live demo,
I'm going to welcome Miles
and Lynn to the stage.
And they are
going to show you
how Agentforce helps
make customer service
as proactive as possible.
So, first, we're going to
start in the Vivint app.
Right here, this is an
AI that you all probably
But what's great is that
Agentforce has Data Cloud
It knows Vivint's
business.
Trusted knowledge
articles,
troubleshooting,
product specs.
So when I ask a question,
it knows what to do.
But let's be
honest, who here
has time to go to a
site, ask a question?
This is a home
security company.
I want them to tell
me that something's
wrong with my
system first.
So let's see
how Agentforce
So we're going to go
right into Data Cloud,
and we are going to
create a new data stream.
Vivint keeps their
telemetry data inside
And I'm sure you all keep
your data in other places.
So we're going to go in
and create a data stream.
And at the top, you'll see
all that Salesforce data
Super easy to connect in.
And as we scroll in, our
200 plus data connectors
And with one click, I
can select Snowflake,
and make some
configurations
Again, we've made
the hard things easy.
And in the background,
what we're doing is
we are taking that
telemetry data
We're doing
identity resolution.
And then we get this
beautiful visualization
At least I think
it's beautiful.
Because I'm an engineer
and I'm a nerd.
But what you
will see here is
This is now connected
into Data Cloud.
But what's more, it's also
connected to the customer.
So all of that information
is right there.
How do we make sure that
it's safe and secure?
Trust is our
number one value.
And so we have built
robust data governance
We are so excited that
we're also using AI here
to make your
job way easier.
When you pull in
all of this data,
AI classification
helps you say,
Maybe there's CCPA
implications, GDPR, HIPAA.
All of us want
to make sure
that our data
is protected.
And so once the
tagging is there,
you can, of
course, edit it.
But when you're ready,
you can create a policy.
So with just a
couple of clicks,
what I'm doing is
I'm determining
when do I apply
data masking?
Who has access to
this information?
But the power
with the platform
is that you build it
once, and it is available
So then this
policy is applied
when you're testing,
when you're in your call
center, on your
website, or it even
helps you decide whether
your legal team has access
to information
or an analyst.
You build it once, you
set it, and you forget it.
And that's how you bring
humans and AI together.
So we have our
policy in place.
Now let's go and
try and test this.
If you look up here,
again, complicated screen.
Because Vivint has a lot
of different products.
And they have
to think about,
for every piece of
telemetry data, what they
would do in terms of
when a sensor is giving
So how might they
use Agentforce
You bring agents
in, of course.
And it becomes
nice and simple.
Because you have that
amazing reasoning engine
of Atlas that I was
just showing you.
I'm here giving
it information
about the customer,
their telemetry data,
And Agentforce can do
that agentic reasoning.
Is this the first
time that I'm
What device
are they using?
Does it have the
latest firmware?
All of that
information is here.
I don't need to
code it, because I
know no one in
this room wants
to keep all of those
flows maintained.
That's not
sustainable either.
You want to really
drive growth.
So now that we have
it there, Miles,
can you trigger
this flow, so we
can show everyone
what agentic reasoning
So what we're going to
see is, I have a device.
It's sort of at
that tipping point
He's going to issue
that debugging flow.
We're applying all
of that governance
And if you look at my
mobile application,
you'll see there's
now an alert.
Because what we are
seeing is that something
And Vivint is able to
provide proactive service.
In the middle,
you can see, oh,
it looks like the
Wi-Fi is dropping.
There's potential issues.
And Agentforce knows that
I have a customer care
And so it's also
recommending a technician
to come to my house,
which is awesome.
And then what
I love best is
all of that telemetry
data, all of that
diagnostics that
I just showed you.
When the technician
comes to my house,
they're also able
to get access
to all of that information
and governance.
So they have all the tools
they need to get started.
And that is how Vivint,
or any of your companies,
can bring proactive
service into Agentforce
So come on that
journey with us.
We're so excited for you.
Give a big hand
to MaryAnn.
You saw how to
build agents
You saw the power
of bringing data,
and unlocking all that
trapped data to power
But now to bring
it all home.
Because it's
also important.
How is that data,
and are those agents,
how is it helping you
with all those customer
touch points across
the Customer 360?
And to tell you more
about that, please
welcome Jon
Moore, director
of product marketing
here in Washington, DC.
It's good to
see you all, DC.
I love looking out
in the audience
and seeing some
familiar faces.
But I recognize,
while some of you
may be familiar with
Salesforce, some of you
may be just
getting introduced.
So I want to share with
you a really grounding
statement about
our mission,
so we're all starting
from the same page.
Since the beginning,
it's been our mission
to help you to create
new connections
with your
customers, connect
with your customers
in whole new ways.
We've done that
by helping you
to build a
Customer 360, which
connects every part
of your business
across the entire
enterprise,
all integrated on our
deeply unified platform.
And now,
Agentforce has been
extended into every single
one of those experiences.
So this is exciting
because we didn't even
have this language in
our lexicon this time
For the first time, we're
landing with agents in DC.
And now every application
has been agentified.
And I'll be able
to show you how
by I just walking you
through the platform,
starting with Sales Cloud.
Agentforce for
sales allows
you to offload a lot of
those tedious, repetitive
tasks to agents,
freeing up your team
to be able to focus on
high-impact work that
really drives the
business forward.
Now, I realize,
I'm starting
with an example that
may not land initially
with everybody
in the room,
because you're not
in the business
Nonprofit,
government, education,
You have a home
here as well.
Because fortunately, as
the world's number one AI
CRM, relationship is
literally our middle name.
And relationships sit at
the heart of Sales Cloud
And those valuable
constituent relationships,
customer relationships,
student relationships
that you own, are
fed and cultivated
by transformative
experiences
that Agentforce
force service
With agents in
service, you've
got humans and
agents together
they're delivering
proactive service
with a personalized touch.
And that helps
customers, like Heathrow,
see a great reduction in
their call handle rate.
And how you reach those
customers and stakeholders
This is your
Engagement Engine.
This is how you are able
to keep everybody involved
And agents help you to
launch highly customized
campaigns, targeted
campaigns that
Time that we just
don't seem to have,
especially if you're
in a fast paced, highly
transactional environment,
like commerce.
Commerce Cloud
Agents give you
assistance on
the buyer front.
You've got buyer agents
helping your buyers
with reordering, personal
shopping agents, which
I could certainly
put to good use,
helping your customers
find the right products
at the right times,
automatically adding
And as you already saw
in Avanthika's demo,
Fisher and Paykel is
using an AI agent, Will,
to help them cut down on
the time they're using,
spending time
working out orders.
And speaking of orders,
with an agentified revenue
and orders, you
can automate
the entire
revenue life cycle
from quoting to
contracting to billing.
And as you heard from
MaryAnn, all of this
Now, Tableau
allows you to be
able to view, interact,
and understand
Imagine being able to
access AI-powered insights
inside of every
application.
That's exactly
what Box is doing.
They're unlocking
powerful insights
and AI-generated
visualizations 97% faster.
I mean, think about
the time and effort
saved multiplied across
your entire enterprise.
Speaking of time
and effort saved,
this is one of
my favorites,
because I live
it every day.
I've said it before,
I'll say it again.
If you're serious
about productivity,
you have to be
serious about Slack.
Slack is a productivity
platform that I use,
and everyone here can
use to bring together
all of your people,
all your data, all
your knowledge, all your
applications in one place.
So you can collaborate
in real-time,
And thanks to MuleSoft
out-of-the-box connectors
and low code tools, now
your agents can take
action everywhere,
including in third-party
Now, all of this has
happened over the course
We launched all
this at Dreamforce.
And we've had to move so
fast, as Parker stated,
because of all the change
that's happened around us.
Think about
how much change
has happened in
your industries
And production is probably
having a meltdown.
I promise, I'm
sticking to the time.
But I do feel
compelled to say this.
Because in the past
several weeks and months,
we've seen so much
change in our city,
It is not an
exaggeration to say
that in the
last six weeks,
we've seen significant
change in the world.
And navigating
transformation
during that
type of change?
That has been
and continues
to be very difficult,
challenging for all of us.
But I do want to state,
whatever talent, whatever
skill, whatever
gift you have
to make an impact
in your industries,
in your part of the world,
you are the change makers.
Technology is not going
to do that for you.
The beauty of
where we are now
is that we're
literally in a place
where technology does
work right beside us,
So wherever you
are, whoever
you are, whatever
industry you represent,
Agentforce is ready to
join you on the journey.
And to help you
get started faster,
we have over 100
industry-specific skills
and actions
for Agentforce.
So what does an agentic
future look like for you?
I work with public sector.
I'll give you a
government example.
For those people who are
responsible for delivering
critical benefits
to families,
what if you had an
agent that could deliver
to you a proactive summary
of benefit eligibility
What would the agentic
future potentially
look like for our
next trailblazer?
That might look
like creating
powerful connections
to help youth unlock
For over 120 years, Big
Brothers Big Sisters
of America has been
changing lives.
And today, they are the
largest youth mentoring
organization
in the nation.
But they're
facing challenges
that are common
to so many of us
in this room,
staffing limitations.
And as a result,
every year,
about 5,000 youth age
out of the program before
being connected
to a mentor.
That is 5,000 missed
opportunities year over
year to potentially change
the trajectory of a young
And so they've
turned to Agentforce
to help their Bigs and
Littles find each other
Take a look at
this short film.
She was shy at the
beginning, like,
when we first
started talking.
We firmly believe
that every child
And it is our
unique opportunity
One person, one
family, one community--
the domino effect of
what we're able to do
We have 30,000 Littles
that are waiting to be
And our goal is
to reach as many
of those as fast
as possible.
The goal isn't just
to make a match.
It's to make
the right match.
That's where Salesforce
was introduced.
Using Nonprofit Cloud,
what we call Matchforce,
all of our matching
specialists
go in and see
exactly where
they are in the
process, all the way
through the entire life
cycle of the relationship.
The efficiency
of Matchforce
helps us double our
output without hiring
Every dollar we
save is a dollar
we can then turn around
and reinvest back
I was thinking of maybe
doing, like, fireworks
And people are
watching from a boat.
Ooh, that would look nice.
We're trying to reach
more young people.
We're trying to
do it faster.
But we also need
to make sure
that the quality
of match increases.
When I heard
about Agentforce,
I thought right away
this could be a solution
Agentforce is integrated
in Matchforce, where
all of this information
on matching criteria
and preferences
already exist.
So Agentforce is helping
streamline our process
and elevate match
recommendations at a rate
that humans just can't do.
Every day matters
that a child
And we expect
with Agentforce
to be able to cut
that time in half.
There is nothing more
rewarding than believing
in children and
then watching
them step up and
believe in themselves.
Even impacting one child's
life is worth the pursuit.
I love being part of Big
Brothers Big Sisters.
Treasure, you're a really
important part of my life.
I just appreciate
you just being here.
As we like to
say, it takes
And with Salesforce,
we are bigger.
That's an
incredible story.
And I'm so pleased
to be able to share
that Big Brothers Big
Sisters of America team
is here with
us in the room.
And one of the stars of
the film you just saw,
their CTO Travis
Gibson is with us.
Welcome all of you to
Agentforce's World Tour
And now we get to see
all of that come to life
Let's welcome back
Miles and Lynn.
Let's give them a
hand really quickly.
And while we
jump in, I just
want to share with you
kind of a grounding
statement about
how we're going
So Big Brothers Big
Sisters of America
One, they want
to ensure they
have a vast network
of volunteers.
Two, they want to make
those matches between Bigs
And three, most
importantly,
they want to make sure
that those matches last.
Now the big
challenge here,
as I mentioned to you, is
that a lot of the youth
are timing or aging
out of the program.
So time is of the essence.
So we're going to attack
this demo in three parts.
And let's start with Vast.
So it's important to
have a lot of volunteer
But keeping our
volunteers engaged
throughout every part of
the onboarding process
And this is where data
makes the difference.
So Tableau
Next, our newest
analytics
experience, gives us
the ability to assess
volunteer retention
And before, what
we'd have to do
is just dig through
manual reports
to try and pull
out these insights.
But now I can ask
conversationally,
why aren't my
volunteers completing
Now, this may not even be
the most ideal question
And Tableau is going
to suggest to me
a few more analytically
correct questions so I
can get to the information
I actually need.
And I'm going to
select that first one
to see where I'm seeing
the highest drop off.
Now, while it's analyzing
this information,
it'll pull back
insights as well as
powerful visualizations
if I want to double click,
We see the
biggest drop offs
are at the interview and
the background checks.
So why don't we go ahead
and make those steps
And I can do that in Flow.
Now what you're
seeing here
is an example of the
volunteer application
And you already saw
in MaryAnn's demo
how we can streamline
these types of flows
So rather than building
an additional conditional
logic, we've already
dropped in an agent that
can start doing
the reasoning,
asking questions
like, when's
the last time this
volunteer was engaged?
What's the best way to
get them re-engaged?
And thanks to
MuleSoft connectors,
we can also
account for steps
that we need from
external systems.
Now that we've updated
the flow on the back end,
let's see how that works
from the perspective
And we're
looking at Chris.
Chris is a
prospective volunteer
with Big Brothers
Big Sisters.
He's completed his
application but not
And we now know that's
one of the big drop
Thanks to our
updated flow,
Agentforce steps in,
detects his inactivity,
sends him a
prompt right here
in WhatsApp,
which is where he
And in one tap, he can
schedule that interview.
No phone calls, no
forms, no back and forth.
That keeps the Big
Brothers Big Sisters
employees and
staff free to focus
on the big things and
not the routine things.
So now we've seen
how Agentforce
is helping with the
onboarding process.
Let's move on to our
second step of fast--
And here we are
in Matchforce.
This is the match
management system
The good news now is that
Chris is officially a Big.
He's completed
all of his steps.
And now we get a
chance to match.
This is probably the
most important part
It's also a good
place for us
to pause and emphasize
the importance
We're talking about
humans and agents.
The reality is our
match specialists
have experienced
insight, intuition
But we're just
making their job
So with Agentforce
working alongside them,
the agent can go through
and analyze reams of data
that normally they
would have to-- all
the interview notes,
all the documentation,
the preferences,
the interests,
and all those
sorts of things--
and then deliver up high
quality recommendations
And at 92%, Ethan
looks like he'll
So let's go ahead and
put them together.
So while Agentforce
is working on that,
let's move on to our final
step, which is where we
start to measure success.
How do we know
we've done well?
Well, did the match last?
We're looking
at Chris again,
So big Brothers
Big Sisters
uses Slack to
keep their Bigs
This is where they
converse with one another.
And this is where they
access valuable resources
And because it's
FERPA compliant,
all their correspondence
is private and secure.
Today, Chris is
using Slack AI
to rehash their last
year and a half together.
And using conversation
summaries,
you can see he's
able to surface
some of their big moments,
celebratory moments
like Ethan's
science fair win,
and even some of the goals
that they established
at the onset of
their relationship,
like improving
Ethan's grades.
And just like
that, they've
hit their 18
month milestone.
And based upon how
things are going,
they're very likely to
have a lot more milestones
Thanks to Agentforce,
Big Brothers Big Sisters
of America is building
a vast network
They're making those
matches fast and with much
And with humans
and agents,
they're creating
life-changing connections
One more time, thank you
so much to Big Brothers
And Parker, back to
you to wrap it up.
All right, big round
of applause for John.
So I love how you really
lit up all the Customer
But I got to say
selfishly, especially
Slack, because
I've been running
engineering for Slack for
a year and a half now.
And if you want to use
agents for employee facing
use cases, I
guarantee you Slack
is going to be the best
place for you to do that.
That's just a
little plug for me.
You saw the Customer 360.
You saw how to build
Agentforce agents.
This AI is, like,
so cool right now.
It's like, oh, it can
do this, it can do this.
And look, I could do this.
We're going to
show-- like, we
show amazing demos,
but start simply.
Just start and build an
agent to answer questions.
That's what we did at
Salesforce on our website.
The first thing
we did is just
A human could
read the website.
They could read all
the product docs.
A lot better experience
to go to the agent
and just ask a natural
language query and get
How does Data Cloud work?
Is it available
in this edition?
All those kind
of questions.
Then move on to ask the
agent to do something
and have the agent--
program the agent
to actually take
action based upon--
an action based
upon a reaction.
And then the scary
part, but it's
move on to
proactive action
where the agent
actually will do work
That is when
it's really cool.
That's really when you
see the power of humans
Well, how are you going
to figure this out?
Like I said earlier
in the show,
you can do this
today for free.
Go online, get a
Developer Edition account.
And if you want to know
how to start learning,
Go to trailhead.com
and become
an Agentblazer Champion--
level 1, level 2, level 3.
We want one million
Agentblazers.
And it starts with each
and every one of you
in the room, in the
overflow, and online.
The show here is not over.
There is a huge
show going on--
110 plus sessions, 25
demos, 30 sponsors.
Go to each and
every one of them.
Or send your agent to go
do some of them for you.
Here in the room, 1:00
PM, there's going to be
a great another keynote
by the public sector
on efficiency and
impact with AI.
I guarantee you that's
going to be really cool--
Go build your first agent.
And most importantly,
give us feedback.
We always aim to improve.
So give us that
constructive feedback.
And thank you so
much for being here.
Enjoy the rest of the day.