Play
Show Navigation
Show Transcript
00:00
Play
Seek 10 seconds backwards
Seek 10 seconds forward
00:00 / 20:36
Mute
Disable Captions
Settings
Picture in picture
Fullscreen
Search in video
You can search the video captions for specific words or phrases.
Popout transcript
Hide Transcript
Skip transcript
MARC BENIOFF:
Now, probably one
of the fastest things that
we've seen really happen
is the growth of this
incredible new generative
AI.
But what I like
to show you today
is this idea of where
our AI capabilities have
gone.
And for those
of you who have
been with us
since 2014, you
saw us first
launch Einstein.
That really came
out of where we all
had kind of an existential
freakout at Salesforce
around 2010.
That's when we
started to think
this AI revolution
was going to happen,
and we had to get going.
We started to acquire a
lot of amazing companies
and bring a lot of
incredible minds
into the company.
And we built this amazing
platform called Einstein.
And that was very
much the beginning
of our AI journey.
Today is another
critical step on it.
But I'd really like
Srini to kind of walk us
through what we've been
doing since we first
introduced Einstein.
So Srini?
SRINI TALLAPRAGADA:
Thank you, Marc.
OK.
This is good.
So I think Einstein
is AI for CRM.
We've been pioneering
AI for CRM since 2014.
And today, Einstein
does about a trillion
predictions every week.
So what does that mean,
trillion predictions?
If you are a Sales
Cloud customer,
if you use our lead
scoring app, Einstein
Lead Scoring, you can
convert your leads
3x faster.
Just in Thanksgiving week,
if you are a Commerce
Cloud customers, we
generate more than 50
billion AI product
recommendations
and customers see a
300 million GMV uplift.
Now, to do this, we
had to first invest
a lot in hiring
world class talent.
We acquired some
great companies
like MetaMind,
RelateIQ, PredictionIO,
and coupled it with a
lot of organic hiring.
So a world class AI team.
Both researchers,
data scientists,
data engineers, and
all and invested a lot.
That's one thing
we had to do.
We also had to solve a lot
of fundamental problems,
because one of the things
that you want to do
in AI, especially in
the predictive case,
was as you're doing
machine learning,
we cannot see the data.
All the data
scientists want
to see a lot of data.
And we hire all of them.
They say, Srini, give me
access to all your data.
I said, no.
You cannot.
So what we have
to do is we
have to invent new
techniques like auto
feature engineering, auto
feature selection, auto
model selection.
And we open source
a lot of these,
but invent new
technologies.
Some of the other
things as we learned
as we are doing it
were as we deeply
invested with
our customers,
new use cases
started coming.
This is when we started
investing in LLMs.
And some of the earlier
papers that we have
invested, like decaNLP.
And all were
suggesting how
prompting could be
a new technology
that we could invent.
And then what we
did is we said,
if this is where
the world is going,
let's invest deeply in
these LLM technologies.
And we learned,
very interestingly,
that some of
these LLMs are
auto-regressive
techniques.
And then we said,
really interestingly,
we could use it for
protein generations.
And then we publish a
lot of protein generation
papers in Nature
magazine and all.
So we started seeing
this early one.
So what we want to do, in
the predictive AI, what
we did is we brought
all these technologies
for our customers.
So we abstract all
of that technology
so that you can
get business wins.
What we are going to
do is in the new world
of generative
AI, we still have
to solve the trust
problems, security
problems, scale problems,
privacy, ethics.
We'll solve all of
that and bring you.
And that's what we
are excited about
to bring you
to the future.
And that's what
today is about.
Thank you, Marc.
MARC BENIOFF: Very good.
All right.
Thank you so much, Srini.
[APPLAUSE]
I want to just stay
on this slide for one
minute.
And I'll tell you, I
don't think a customer
interaction or story
or meeting today
starts or ends
without this idea
of what's really happening
with generative AI.
And the story usually
goes like this.
I know about these new
large language models.
Many of these
models are amazing.
A lot of us have used
ChatGPT and this kind
of GPT-4 model.
Also a lot of
the models that
have been created by our
own engineering teams
are amazing.
And many of the startups
who are in this room
are also building
these amazing models.
And so these CEOs are
so inspired or CIOs
that we're talking to.
And they're
like, what we're
going to do is we're
going to start working
with these
models, and we're
going to take all of
our corporate data.
We're going to put
it into the LM.
And then all of
a sudden, we're
going to have instant,
intelligent company.
And it's not
quite like that.
OK.
It's not quite like that.
Now, the reason
why they think that
is, they already know how
these other more public
models are working.
They take these huge
vacuum cleaners,
and they're just vacuuming
up all of the data
off the internet
that they can get.
And you see that.
So they're all
just taking--
if there's publicly
available data or just
data that's out there
off of an internet
that can be
scraped, they're
taking it and training
their models, which
is much data as
they can bring down.
And then they're
taking that data,
and then they're
turning on their LLM.
And whatever comes
out of it is great.
And if there is this
concept of hallucination
or if the LLMs,
basically, starts to lie,
well, that's not really
their responsibility.
Their responsibility
is to give you
the best case they can
with their generative
artificial intelligence.
That's not exactly
the world of trust
that we live in.
And recently, I was
with one of our largest
customers right
here in New York,
very large bank.
And the CEO made
it very clear
that they want to use
LLMs to become much more
productive.
In mortgages, and
account service,
and all the capabilities
that a bank will do.
So can they do that?
Can they just take all of
their account information
and all of the
history of all
the accounts
and everything,
just put it into an LLM.
Well, I don't think
that's going to work out
very well when it comes to
the regulated industries
in the way that data
works in large companies.
That's why really,
the onus becomes on us
to give our customers
this next generation
of generative AI.
And that's what you're
going to see today.
Because, as I
mentioned earlier,
and this is going to
ground our presentation
today, trust is our number
1 value at Salesforce.
We saw that when
Parker and I first
designed the
product, which
was way back in 1999, we
came up with something
called the Sharing Model.
And that idea was
that every person
in the company who gets
Salesforce access, they
also get access
around what data they
can see or not see or use.
Because we know,
especially like
for a bank, every bank
executive cannot see what
every other bank
executive can see.
Or in health care.
What one health care
executive can see,
another health care
executive can see.
We all understand the
enterprise sharing model.
Don't we?
All understand what
that means in a company.
And we also
understand that a lot
of us in
enterprise IT come
from the relational model.
And the relational model
is the rows and columns
model of data where
it's security is down
to the cell level.
Not only where readers
don't block writers,
but the idea that cells
can be locked by one
particular user so that
cell cannot be seen
by another user.
That is not the kind
of generative AI
that we're all
experiencing.
Is it?
Because the way these
large language models
work is they're taking
down all that data
that they grab,
and then they're
amalgamating and
tokenizing that data
and then using
their algorithms
then to generate
their intelligence.
So that idea of
cell-based security
is not in the current
generative model.
That's the breakthrough
that we're really
going to try to show you
today where we really
are leveraging
starting back in 2016,
where we came up with
our first trust model
for predictive.
Now, Parker and I were
joking about this today,
because I always called
it anonymous predictions.
Because the idea
is that when
our systems, when our
applications, when
our platform looks at
all of your data and then
uses machine
intelligence or machine
learning or deep learning,
which are the three
primary artificial
intelligence
techniques that
really existed
before the current
generative AI techniques,
we don't look
at your data.
We're able to provide
you those predictions
and that AI capability
without actually looking
inside the data by just
keeping it anonymous.
And now, with generative
AI, what we're able to do
is we're able to take
the same technology
and the same idea to
create what we call a GPT
trust layer, which
you're going to see today
for the first time.
And we're about
to roll out
to all of our customers.
So they have the ability
to use generative AI
without sacrificing their
data privacy and data
security.
This is critical for
each and every one
of our customers
all over the world.
Every transaction and
every conversation
at Salesforce begins and
ends with the word trust.
So we understand
that very well.
And there's one
other critical part
of all of this.
It's not just
about trusted AI
and delivering
the technology
to the right person
at the right time.
But it's also about
responsibility.
As we're all
going to learn,
because we're now all on
a societal AI journey,
there is going to be a
lot about responsibility
in this technology.
We've all seen the movies.
And we've all seen
where this can go.
Haven't we?
We all have
these crazy ideas
in our head of
what could happen.
There's many different
possible scenarios.
So that's why responsible
AI use is so critical
and why I'm so
excited that we have
an incredible AI
ethics team that
has been in place
at Salesforce
for many years.
And Kathy Baxter
is here who
wrote an incredible
article that
got published in HBR this
week on ethics in AI.
But I asked her to
speak on our panel.
And Kathy, would you
just ground us right now
in responsibility
of AI before we
begin our technical
journey today?
KATHY BAXTER:
Thank you so much.
MARC BENIOFF: Thank you.
KATHY BAXTER: In 2019, we
published our trusted AI
principles.
And the very first
principle is responsible.
We believe that
at the core,
our responsibility
is to ensure
that we are protecting
human rights, as well
as the data that all
of you, our customers,
entrust us with.
And this first line, your
data is not our product.
That is a key
differentiator.
Your data is
not our product.
At the beginning
of this year,
as we began delving even
more into generative AI,
we recognized
that we needed
more specific guidelines
than the principles that
have been driving
all of our work.
And so we created
these five guidelines
that are specific to
not just generative
AI, but generative
AI in an enterprise
environment.
First and foremost,
it has to be accurate.
You're making
business decisions
based on the content
that's being generated.
So it must be accurate.
It also has to be safe.
We need to assess it
for bias and toxicity.
It needs to be
transparent.
We are ensuring that the
data provenance of all
of the data that
trains our models,
it's fully
consented and we
are transparent whenever
content is AI-generated.
We believe that AI
needs to be empowering.
So keeping humans
in control.
Giving them the tools
that they need to know,
is this content
useful, is it accurate?
And then finally,
sustainability
is one of our core values.
And so when we are
building models,
we are going to
ensure that they
are right sized.
It's not about
creating the biggest
model possible, it's
creating the right size
model that is
going to be most
accurate for
our customers,
but also takes
into consideration
the carbon and
water footprint
necessary to train them.
Thank you all very.
Much.
MARC BENIOFF: Very good.
Thank you very much.
[APPLAUSE]
We're so lucky
to have Kathy
and also our
incredible ethics team.
And guiding
us, I think, is
going to be a huge
burden for them
as we start to move
forward very rapidly
in this new AI journey.
We already
understand this is
going to be a
critical part
of every single
thing that we do.
In the world that
we live in and in
these industries
that we participate
in like banking or health
care or media or so many
of the industries that are
here in this room with us
and all watching us
all over the world,
they like to know exactly
where their data is.
And they want to
know where that is
and audit that place.
And that is not how
generative AI works.
Generative AI works with
this kind of expansion
from the deep learning
principles where
deep learning have these
amazing neural networks
and have these
amazing insights.
And now the
networks just got
expanded and expanded.
And as the
network expanded,
all of a sudden, the
ability of the models
expanded as well.
But as those expansions
happen, the weights
and how they're
moving all are
shifting and changing.
And that is
really the burden
of our AI team over here.
They have to really
be able to use
these next
generation models
but have that capability
to deliver a trusted
experience to
our customer.
We're already
the number 1 CRM.
We already know that.
All of our
customers are doing
a trillion transactions
a week using Einstein.
There's no company that's
even close to what we're
doing in the customer
relationship management
area with artificial
intelligence.
We understand
the burden there
for must be on
us as we're going
to take this forward.
We're already
trying to do all
these incredible things.
To maximize return
on investment
for our customers, to
deliver them the fastest
time to value, to
innervate the fastest
with low code or no
code capabilities.
But now, giving them
the trusted productivity
that they're
demanding with
generative artificial
intelligence.
How will they augment the
employees productivity
just as we heard from
our friends at Gucci?
Now, our taste of
that has really
started to accelerate
earlier last year when
we introduced
our customers
to this incredible
new product
that we have
called Data Cloud.
And Data Cloud has become
our fastest growing cloud
ever.
And one of the reasons
why this is becoming such
an important cloud
for our customers
is, as every
customer is preparing
for generative
AI, they must
get their data together.
They must organize and
prepare their data.
So creating a data
cloud is so important.
But the problem for a
lot of our customers
is that they might be
creating data clouds
but with teams or
with technologies
that are outside of the
Salesforce ecosystem.
And that's why we extended
our Salesforce core
platform with
this product that
was intelligent,
real-time, automated,
and hyperscale.
We introduced it at
Dreamforce last year.
It's taken off
like a rocket ship.
And we've seen this
incredible capability
now.
We're already delivering
$30 trillion transactions
per month with Data Cloud.
And we're
importing, I think,
about $12 trillion
components
of data already on an
fully annualized basis.
It's incredible what
our customers are doing.
I mean, one of the amazing
stories that we have.
We work so closely
now with Ford
and these next
generation cars
that they're
building and trucks.
I have one of these
new lightnings.
They're amazing.
It's just
dripping telemetry
and it drips data because
of all the technology
and all the capability.
But it needs a
receptacle for that data.
It needs the ability
to have all the data so
that Ford can provide
the next generation sales
experience.
I just bought a
second truck actually,
because it was so amazing.
A service experience,
a marketing experience.
All of those things
are then generated out
of the Data Cloud.
So the Data Cloud
sets the stage
and the beginning for
every customer's AI
journey.
And then our
primary vision,
the vision that
Parker and I set
up with 24
years ago, we're
ready then to
begin connecting
with our customers
in a whole new way.
This idea that we
want to then provide
that customer 360
experience from sales,
to service, to marketing,
to commerce, to Tableau,
to Slack, to all
the capabilities
of Salesforce on
the 360, but then
augmented by
generative AI.
We already know
generative AI
is changing everything.
We already can see
the huge growth.
How many people here have
already used ChatGPT?
Raise your hands.
Everybody.
So I didn't have
to call for hands.
Did I?
It's the fastest
growing consumer product
of all time.
We all know the
story, and it's
an amazing experience.
But we've all seen
the limits as well.
And we all know that
all the data that we're
putting in there,
as we sit there
at home and play things
or whatever, well, it's
getting stored and you're
training that model.
And that's how it works.
And if you want to
understand that,
talk to some of
the AI experts
in the room today, and
they'll explain to you
why that consumer model
is so powerful for OpenAI.
Of course.
But does that work
in our enterprise?
Does that give us that
trusted experience
in our enterprise?
Is it going to give us
trusted productivity?
We want the augmentation.
We want that next
generation capability
for our enterprise.
But what about the trust?
How are we going
to store that data
when what the LLM's
desire is to take as much
of that data and then put
it into its weights as
needed.
But is that
what we're going
to allow it to do with all
of our enterprise data?
Will we be able to
preserve our sharing
model?
Will we be able to
preserve our security
model?
Our privacy model?
Will we be able
to lock down
for each and
every customer
and each and every
employee, what they need?
That's why we already
know that every CEO needs
an AI strategy now.
We've been talking about
that for over a decade.
And we have the
best team when it
comes to enterprise AI.
There's no question.
But when we look
at these new models
that everyone's
going to roll out,
we know one thing.
And that is, there's
a pretty big gap.
And it takes place
in every conversation
I have with every CEO.
It's this I trust gap.
When we talk
about privacy,
when we talk about
hallucinations,
when we talk about
data control,
when we talk
about bias, when
we talk about toxicity.
These are technical
terms, actually, in AI.
They're not like kind
of societal terms.
These are actual
technical explanations
of things that
are happening
inside these models.
That we realize
that over here,
we have this desire to
rapidly move forward,
to have more productivity.
But over here, we have
this need for trust.
And what we
hope to do today
is to really
close that gap.
This is a critical time.
That's why I
mentioned, we already
have delivered that first
generation of our trust
layer with our
predictive model in 2016
when we rolled
out Einstein.
We already crossed
that bridge.
And every customer who's
in this room and online
who uses Einstein
has already
crossed that bridge.
We never had to say
to anyone, oh, yes.
We had to look
at all your data
to make a prediction.
No.
We didn't have to do that.
And we also know
that we have
to apply it to all
of our applications,
not just some of them.
And we also need
to then say,
we're going to
take all of this
to not only be number 1
AI CRM in predictions,
but number a
AI CRM when it
comes to this incredible
generative AI capability.
So that is why we're
introducing today
our AI cloud.
Trusted enterprise
AI built for CRM,
built for Salesforce,
built for our customers.
Implementing these
key technologies that
are so critical for us.
Using our Einstein
GPT trust layer,
which is about
to get rolled out
for all of our
customers worldwide.
Allowing all of
our applications
to deploy that capability.
And augmenting our
own capabilities
to be the number
one AI CRM.
Resume AutoScroll
Questions about Salesforce+? We have answers.
What is Salesforce+?
Salesforce+ is the free streaming service for business. It’s a place to watch live events and original series that help you grow your career and company, on your own time. Salesforce+ is produced by Salesforce, so you’ll find our company’s big events like Dreamforce here — as well as new original shows and virtual learning made for your role and industry. Everything is free and unlimited.
What can I watch on Salesforce+?
You can watch all kinds of helpful, entertaining, and inspiring videos on Salesforce+. One thing our members love is our events. Whether you catch them live or on demand, Salesforce+ gives you a free front-row seat to our biggest global events, like Dreamforce, TrailblazerDX, and Connections. Our members are also big fans of our original series. These entertaining business shows star people in roles and industries just like yours, giving you fresh ideas for how to grow your company and do better work. The best part is, it’s all free.
Do I have to pay for Salesforce+?
Salesforce+ is free. Watch any Salesforce+ events or series at no cost. To enjoy unlimited free access to our entire library of videos, become a member by signing up for Salesforce+ with a Trailblazer account. You don’t need a Trailblazer account to watch original series. To join us for virtual events like Dreamforce or Connections, you need to create your own Trailblazer account (it’s free!).
What is a Trailblazer account?
Your Trailblazer account gives you access to a unified profile across the Salesforce ecosystem and allows you to manage your settings and access across multiple Salesforce sites.
How do I sign up to register for live and on-demand events, like Dreamforce?
Sign up or log in to your Trailblazer account and complete a few additional details to watch Salesforce+ events live and on demand. We recommend preregistering for all events on Salesforce+ to get your own personalized journey, reminders, recommendations, and more.
Do I need to be a Salesforce customer to watch Salesforce+?
You don’t need to be a Salesforce customer to watch Salesforce+. Our virtual learning is open to everyone.
