00:00:00
- HubSpot turned 18
years old this past June
00:00:05
and since then I've had the
honor of sharing a number
00:00:08
of memorable moments on the INBOUND stage
00:00:11
with you over the years.
00:00:14
13 years ago, I shared a
launch of my son Sohan.
00:00:19
Back then he was a tiny language model.
00:00:24
Now he's a middle schooler,
00:00:26
so he has a teenage language model,
00:00:29
which I will confess I have
not completely decoded yet.
00:00:32
Parents in the room know
what I'm talking about.
00:00:34
Now I can't show a current image of Sohan
00:00:37
because he's revoked my licensing
rights to use his image.
00:00:41
No cap. No cap.
00:00:46
10 years ago we launched the HubSpot CRM.
00:00:51
Many of you in the audience were here.
00:00:52
Applaud if you were here for
the launch of inbound CRM. Yep.
00:00:55
My lovely wife Kirsten, she was here. Yep.
00:00:58
By the way, she's the one that got me the
00:00:59
rocking shoes just for you.
00:01:00
I know you can't see them on
stage. You'll see them later.
00:01:02
Yeah. Thank you honey. So one
00:01:07
year ago we launched our first
products with generative ai,
00:01:11
including ChatSpot. ChatSpot
started as a Labs project
00:01:16
and now has graduated to
become Breeze Copilot.
00:01:20
They grow up so quickly. So,
00:01:25
this is the year of AI agents.
00:01:27
When we look back at this event
years from now, you're going
00:01:30
to remember it as the year of AI agents.
00:01:34
Now agents are all the rage this year,
00:01:37
but you know what was the rage in the 1980s?
00:01:38
Corduroy pillows.
00:01:44
They were making headlines everywhere.
00:01:48
I'm gonna give that one
a couple of seconds just,
00:01:51
okay, So here's what I'd like to do.
00:01:55
First I'll give you a debrief of
00:01:57
what has happened in the world
of AI since we last spoke.
00:02:02
Then I'll try and demystify AI agents
00:02:05
because they can be
unnecessarily mysterious.
00:02:09
And finally, I'll show you some software
00:02:11
that's in development
00:02:13
and as is permitted in
my speaker's contract.
00:02:17
I'll sneak in some dad jokes.
I apologize for nothing.
00:02:22
I bring my full dad self to the stage.
00:02:25
Speaking of dad jokes, so I'm
00:02:27
building a dad generator agent.
00:02:30
This is true story and
I collect all of them.
00:02:32
I've got 2000 in a,
what I call a dad-a-base.
00:02:38
And so this is basically
feedback for the training agents.
00:02:43
So thank you for laughing for those
00:02:44
that laugh as we go along.
00:02:45
But anyway, so here we go.
00:02:47
So a lot has happened in the past year,
00:02:50
all focused on just the
highlights, things that matter
00:02:52
to you and your company.
00:02:55
First, AI models are getting bigger.
00:02:59
Bigger in terms of the number
of parameters which sort
00:03:01
of defines how big
their neural network is,
00:03:04
and also their context
windows, which defines
00:03:07
how much information we can
pass to the model at the time
00:03:10
we give it a prompt. Bigger models are
00:03:13
generally more capable.
00:03:15
They can reason better,
they can write better,
00:03:19
they can follow instructions better.
00:03:22
But models aren't just getting better.
00:03:25
They're getting better, faster.
00:03:28
So years ago, Gordon
Moore, the CEO of Intel,
00:03:32
coined what's called Moore's Law.
00:03:36
It states that computers
double in power every 18 months.
00:03:40
That's an exponential curve.
00:03:42
That's why a computer that used
00:03:43
to fill a room now fits in your pocket.
00:03:47
So Moore said: computers would
double in capacity every
00:03:50
18 months.
00:03:52
And AI came along and said, hold my beer.
00:03:57
AI models are doubling every six months.
00:04:01
And this growth is partly
what's causing all this
00:04:04
excitement around AI.
00:04:07
And it's not just that the
models are getting bigger
00:04:10
and better, but we're
seeing innovation along
00:04:13
multiple dimensions.
00:04:14
We now have multiple highly
functional frontier models
00:04:18
that are good at different tasks.
00:04:21
Open AI's. GPT is really
good at reasoning.
00:04:25
Claude 3.5 is really great at writing.
00:04:29
Google Gemini is really great
at supporting huge context
00:04:33
windows with millions of tokens.
00:04:36
And Llama is really great
at being open source
00:04:40
and having a cute name.
00:04:46
And we have some breaking
news on the AI model front,
00:04:49
for almost a year now,
the industry has been buzz
00:04:53
with the anticipation of
a new model from OpenAI.
00:04:57
Those are the folks that make chat. GPT.
00:04:59
The model's code named
Strawberry. Word on the street was
00:05:04
that it was going to be very impressive.
00:05:07
What's not very impressive is that joke.
00:05:09
I did not write that joke.
I will not take credit.
00:05:14
OpenAI launched that new
model this past Thursday
00:05:19
and I've been up late
nights even later than
00:05:21
usual every night since.
00:05:23
The model is called GPT o1.
00:05:26
Now the reason to be
excited about GPT o1 is
00:05:30
because of its ability to reason.
00:05:33
It's shockingly good,
00:05:34
like graduate student PhD level reasoning.
00:05:37
And it even shares the kind of chain
00:05:40
of thought it had in
terms of how it came up
00:05:42
with the response that it gave you.
00:05:43
So it sort of lets you
inside its brain. Super cool.
00:05:47
Now this level of reasoning
means we can have agents now,
00:05:50
they take on much more sophisticated goals
00:05:53
that when we could even do a week ago.
00:05:55
That's how fast this stuff is moving.
00:05:58
So the lesson here is that when
things are moving this fast,
00:06:01
we shouldn't go to where the AI is.
00:06:05
We should go to where AI is
going, hat tip to Wayne Gretzky.
00:06:10
And we're going to be in a
very different place three
00:06:13
months, six months, 12 months from now.
00:06:16
Okay? So we have a
variety of great models.
00:06:19
They're also becoming
increasingly multimodal.
00:06:22
So we've moved beyond text to
the models supporting images,
00:06:26
audio and video too.
00:06:29
This means the models
can take image as input,
00:06:32
not just the prompts that we give it.
00:06:34
This is like we're giving
these AI models the
00:06:36
sense of sight.
00:06:38
That's like when you're
teenager can first start seeing
00:06:40
the clothes on their bedroom floor.
00:06:41
It's amazing when it
happens. It's amazing.
00:06:46
So this is what AI models used
00:06:48
to see when you sent them to a webpage.
00:06:50
This is the webpage for School of Rock,
00:06:53
longstanding HubSpot customer.
00:06:55
And I'm the parent of a recent customer
00:06:59
of School of Rock.
00:07:00
So this is what the models used to see
00:07:04
and this is what the models can see today.
00:07:07
They see the same webpage as we do.
00:07:13
So quick recap of the debriefing.
00:07:15
Models are getting bigger and better.
00:07:18
We have different models
for a diversity of tasks.
00:07:22
And AI models now have multimodal support.
00:07:25
This makes it possible for the first time,
00:07:27
to create AI agents
00:07:29
that can take on a wide
variety of assignments.
00:07:34
But agents can seem a bit
mysterious and complicated.
00:07:36
So let's try to demystify them.
00:07:40
Let's raise your A-IQ now,
00:07:43
agent IQ speaking IQ,
00:07:46
I read somewhere that
talking to yourself is a sign
00:07:51
of intelligence.
00:07:53
I've said that to myself
hundreds of times.
00:07:58
Now, first of all, you
might be wondering like
00:08:02
what is an AI agent anyway?
00:08:04
And there are a lot of
definitions out there.
00:08:08
The one I prefer is the one Yamini shared.
00:08:11
An AI agent is software that uses AI
00:08:15
and tools to accomplish a goal
that requires multiple steps.
00:08:21
That's it. Yes, some
agents can have be able
00:08:25
to run autonomously,
00:08:27
some have executive planning capabilities,
00:08:29
but those are niceties, not
necessities to be an AI agent.
00:08:35
And there are a range of agents in terms
00:08:37
of the complexity of the goals.
00:08:38
They can take on. Everything
from complex goals like the four
00:08:41
agents that Andy just
demonstrated, all the way down
00:08:44
to very simple goals.
00:08:46
The thing I want you to
remember is that an agent's
00:08:48
and agent, no matter how small
00:08:54
now, whether big or small,
00:08:56
all agents share some things in common.
00:09:00
An agent uses one or more AI models.
00:09:03
If it doesn't use AI, you can't
officially call it an agent,
00:09:08
then it's just sparkling
automation software.
00:09:13
Thank you. Yeah, bye. Next,
00:09:16
an AI agent has the ability to use tools
00:09:19
so it can access things
like the HubSpot Smart CRM
00:09:22
and take action on your behalf.
00:09:25
And finally, agents have a memory.
00:09:28
They can remember things across tasks.
00:09:31
Otherwise it's like me
walking up to the refrigerator,
00:09:33
opening it and then forgetting
00:09:34
why I opened the refrigerator.
00:09:38
Alright, so let's dig in
00:09:40
and talk about what's in development.
00:09:43
I'm gonna share with you
what I've been cooking on
00:09:47
late at night, two o'clock
in the morning every night.
00:09:49
My wife can attest to this.
And by cooking I mean coding.
00:09:54
I can't actually cook,
not in a literal sense.
00:09:57
And I also don't stand
at my standing desk,
00:10:00
which doesn't sit well
with my personal trainer.
00:10:04
I actually don't have a personal trainer.
00:10:05
I used GPT. By the way,
00:10:09
it suggested that I maybe do like lunges
00:10:13
and I think that would
be a big step forward.
00:10:18
And so having the honor of
being up on stage at INBOUND
00:10:22
and sharing the stuff I've
been working on is the favorite
00:10:26
time of my whole year.
00:10:27
And I've probably spent 1500
hours working on the next 15
00:10:31
minutes that I'm about to show you.
00:10:33
And before I do that,
00:10:35
before we dig in, a quick
disclosure and warning.
00:10:39
The upcoming segment is rated X,
00:10:44
X for experimental.
00:10:48
Everything Andy showed you
are part of real products
00:10:51
that we sell at HubSpot.
00:10:54
Products you should buy.
00:10:55
By the way, if sales team is hiring, I'm,
00:10:57
that's my sales pitch right there.
00:11:00
What I'm about to show
you is an experiment.
00:11:01
Viewer discretion is advised.
00:11:04
And the product I've been
staying up late every night
00:11:07
working on for many moons
this year is agent.ai.
00:11:12
Thank you. Agent.AI
00:11:17
is the number one professional
network for AI agents.
00:11:22
It's also the only professional network
00:11:26
for AI agents.
00:11:29
Now at this point, some
00:11:30
of you might be thinking, wait, what?
00:11:34
Why do agents need their own
00:11:37
professional network?
00:11:40
Reasonable question. 'cause I predict
00:11:43
that in the future our teams
are going to be hybrid.
00:11:48
They, they'll consist of humans like you
00:11:52
and AI agents like me.
00:11:54
Just, just kidding. I'm not
an AI agent yet here on stage,
00:11:58
that likely won't happen
until far into future.
00:12:01
Like INBOUND 25. By the way, inspired
00:12:06
by Andy's references to back the future,
00:12:09
I was going to tell you
guys a joke about time
00:12:11
travel, but you didn't like it.
00:12:17
So here's the homepage for agent.ai.
00:12:19
We see a listing for some
of the agents on the network
00:12:23
and every agent has a profile.
00:12:24
So we can click into the profile,
we can follow that agent,
00:12:28
we can hire that agent and start using it.
00:12:30
It's free. And we have,
00:12:33
and we can add the agent to our team.
00:12:35
And we have a bunch of agents
around marketing, sales,
00:12:38
customer service, operations,
00:12:39
and a new agent's being added every week.
00:12:42
So I invite you to check it out.
00:12:43
You can sign up for free, get a
00:12:44
hundred credits, it's awesome.
00:12:47
So it's time to meet some of these agents.
00:12:51
Now we're gonna look at agents
that help us with marketing,
00:12:55
selling and how to delight customers.
00:13:00
Now the number one challenge
00:13:01
for marketers today is generating traffic
00:13:05
and qualified leads.
00:13:08
I'm not making that up.
00:13:09
One of the most clueful people I
know in our industry said this
00:13:13
circa 43 minutes ago.
00:13:18
So it's even more important
that we take the website traffic
00:13:21
that we have coming right
now, those website visitors
00:13:23
and convert them to leads at
the highest rate possible.
00:13:28
And the way to do that is with
conversion rate optimization.
00:13:32
Now, here's the website for School
00:13:33
of Rock that we saw earlier.
00:13:35
It's a great website.
00:13:37
Is there anything that can
be tweaked or improved?
00:13:41
Yes, always.
00:13:43
But is it a pain in the butt
to figure out what to change?
00:13:47
Yes, always.
00:13:50
'cause you have to review the copy,
00:13:53
you've gotta check the CTA,
00:13:54
you've gotta tweak the headlines,
review the testimonials.
00:13:56
There's just so much to do.
00:13:58
And conversion rate optimization is a very
00:14:00
specialized skill.
00:14:03
Thankfully, we've reduced this
down to a few easy clicks.
00:14:08
Here's the conversion
rate optimizer agent.
00:14:11
It's super easy to get started.
00:14:12
All you do is you type in the web link
00:14:15
for a webpage, any public webpage.
00:14:16
It can be your homepage,
it can be a landing page,
00:14:18
it can be your pricing page.
00:14:20
And the agent goes through and
looks at that entire webpage
00:14:22
and it has decades of domain expertise
00:14:24
and conversion rate optimization
00:14:26
and will come up with
a list of suggestions.
00:14:28
It's like you're saying something
this way on the webpage,
00:14:31
I'll bet you you'll get higher
conversions if you change it
00:14:33
to this, or you're missing a call to action,
00:14:35
or maybe you should add imagery.
00:14:37
It comes up with really,
really good suggestions.
00:14:39
I encourage you to try it.
So it's, it's a simple agent,
00:14:43
but it's a profitable one.
00:14:44
I encourage you to check it
out. Now, Yamini also mentioned
00:14:48
how the average sales rep
spends only two hours a day
00:14:52
with customers and a large part
00:14:54
of their time is spent in prep,
00:14:57
and a large part of
their prep time is spent
00:14:59
researching companies.
00:15:02
Now, the reason this takes so
much time is you have to go
00:15:05
to Google, you go to the
company's website, you go
00:15:06
to Crunchbase and look up their funding.
00:15:07
You go to YouTube, go to the
social media channel, you go
00:15:09
to Glassdoor, you're jumping
all over the internet on this
00:15:11
adventure hunt with all
this unstructured data.
00:15:15
We built a company research
agent that picks up all
00:15:18
of those pieces all over the internet
00:15:20
and puts it together for you
with just a few easy clicks.
00:15:26
Let's look at how easy this is.
00:15:28
We start by entering the website's domain.
00:15:31
I'm going to enter openai.com,
my favorite AI company,
00:15:36
and we get a beautiful multi-page report
00:15:39
broken out into sections.
00:15:41
So we can click on any of these sections.
00:15:43
So let's say we click on founders to see
00:15:45
who the founders of the company are.
00:15:47
We can click on the new section
00:15:49
and see if there's anything new
00:15:50
that's been published about the company.
00:15:52
We can click on their
web traffic information
00:15:54
or their organic keywords.
00:15:56
A bunch of different sections,
00:15:57
brings it all together in one place.
00:15:59
But the coolest part is this
research question section.
00:16:03
You can type in any question you want.
00:16:05
It's like, let's say you're
selling payment software
00:16:06
and you wanna know whether
this company sells a
00:16:08
subscription product.
00:16:10
The agent will crawl the
company's entire website,
00:16:12
crawl the internet, and come
up with an answer for you.
00:16:14
And it says, yes, OpenAI
sells a subscription product.
00:16:18
But what's really cool is you can add
00:16:19
that question to the agent.
00:16:21
So every future report
00:16:22
that you run will automatically answer all
00:16:25
your custom questions.
00:16:26
So let's say yes, OpenAI is a
good fit with a single click.
00:16:31
We can add this company
to the HubSpot Smart CRM,
00:16:36
and we can see that company record
00:16:38
and can do all the things you can do
00:16:39
inside HubSpot's smart CRM.
00:16:41
And finally, let's say we're
interested in this company,
00:16:44
we can click the track button
00:16:46
and type in English what we want to track.
00:16:48
It's like, okay, if they
launch a new product,
00:16:50
if they raise a new round of funding,
00:16:52
if there's an exec change,
describe it in English,
00:16:55
whatever is you're interested in tracking
00:16:57
and now this agent works on your behalf,
00:16:59
and will track all of those
things automatically for you
00:17:02
and notify you if something
interesting to you happens.
00:17:06
So that's the company
research agent, by far
00:17:08
the most popular agent
on the network right now.
00:17:12
Thank you. Yep.
00:17:19
Now what if I told you agents can
00:17:23
use other agents?
00:17:25
I'm telling you, agents
can use other agents.
00:17:29
We can use existing
agents as building blocks
00:17:31
for building new, more powerful agents.
00:17:34
It's like agent composition.
00:17:36
I haven't had this much fun
since I played with Lego.
00:17:39
So this weekend, basically. Okay,
00:17:43
so let's say I'm a
customer success manager
00:17:47
and Canva is one of my accounts.
00:17:50
Love Canva, by the way.
Shout out to Canva folks.
00:17:53
Yeah, I know. Wonderful company.
00:17:54
So now as a customer success manager,
00:17:56
before hopping on a call, I'd like
00:17:59
to kinda update myself
what's been going on
00:18:01
with this company, both
within the HubSpot Smart CRM,
00:18:03
but then all over the web, news, funding.
00:18:06
Now, this sounds familiar,
a lot of that work,
00:18:07
the company research agent already does.
00:18:10
So we built a simple agent that takes all
00:18:13
of this information and
reduces all that work in terms
00:18:17
of creating an executive
briefing into a few easy clicks.
00:18:21
Let's see how easy this is.
00:18:22
So here's the executive briefing agent.
00:18:24
It knows which company I'm
going, which one's assigned
00:18:27
to me, and it will come
00:18:28
and it will say, okay,
I'm gonna look at the CRM,
00:18:30
I'm gonna look at the last calls.
00:18:32
Any issues, I'm gonna go look
at the customer research agent
00:18:34
and I'm gonna pull it all
together into a very brief
00:18:38
executive briefing
00:18:39
and I'm gonna convert it to an audio file
00:18:42
so you can just kinda
listen to it on the commute
00:18:43
to work if you want to. Right,
00:18:44
just pulling all that information together
00:18:46
and giving it in the form that you want.
00:18:49
Alright, so given the agents
that Andy has demonstrated,
00:18:53
given the agents I just showed you
00:18:55
and the two dozen other
agents on the agent.ai network
00:18:58
already, and more coming every week,
00:19:00
I could make the prediction that most
00:19:02
of you will be using agents this year,
00:19:05
and that's an easy prediction to make.
00:19:07
So I'm going to make that prediction.
00:19:08
Most of you will be
using agents this year,
00:19:11
but I'm not going to stop there.
00:19:15
I predict that many
00:19:16
of you will be building agents this year
00:19:21
and not just the introverted
developers like me.
00:19:25
Anyone can build an AI agent.
00:19:28
You can build an agent,
you can build an agent.
00:19:30
You can build an agent. Yes.
00:19:33
By the way, if you know,
00:19:35
I could have my like Steve
Balmer chant right now,
00:19:37
it's like agent builder,
agent builder, agent builder.
00:19:40
But I do not have the dancing
skills of my friend Brian.
00:19:45
So we'll just enjoy him for a few seconds.
00:19:47
This is him on the inbound
stage years ago. Hi Brian.
00:19:51
By the way, thankfully Brian
does not drive a hard bargain
00:19:54
in terms of using his
image in my presentations,
00:19:57
which is I'm grateful for.
00:20:01
So announcing Agent Builder,
00:20:04
which lets you build agents.
00:20:07
Why should I have all
the fun? So let me show
00:20:11
you how easy it is.
00:20:14
So here's the agent builder for
the agent I just showed you,
00:20:17
the customer briefing agent.
00:20:19
So it goes through five steps.
00:20:20
It says, okay, go talk to the HubSpot CRM
00:20:22
and figure out which account
has been assigned to me.
00:20:25
Now go to the research agent,
00:20:27
collect all of that information.
00:20:29
Now go pass all of that context
00:20:32
to a large language model,
generate an executive briefing,
00:20:36
and then generate the
audio file so I can listen
00:20:38
to it on my iPod or my music player.
00:20:41
And that's it. It's drag and drop.
00:20:43
So I'll give you a sense of,
if we click on the plus button,
00:20:45
there are other actions we could add.
00:20:46
Let's say we wanted to
add another step to this,
00:20:48
and we have access to all sorts of data
00:20:50
inside the HubSpot CRM.
00:20:51
We can build UI, we can
do all sorts of things.
00:20:54
It's awesome. So the coolest
part about Agent Builder, it has
00:20:58
access to all the frontier models.
00:21:00
So it has access to GPT
to Claude, it has access
00:21:03
to all the image generation models.
00:21:04
Dall-E 3, Flux.1, Ideogram, Playground, name it.
00:21:08
It has access to all the HubSpots.
00:21:09
CRM has access to people
to create blog posts,
00:21:12
create webpage, has Google, has YouTube,
00:21:14
Twitter, everything.
00:21:15
You can mix it all up and
build an agent with drag
00:21:18
and drop in the same way you
build workflows in a marketing
00:21:20
automation product like HubSpot.
00:21:23
Alright, I'm gonna take a breath.
00:21:25
This is a cue for me to breathe
00:21:26
so I don't pass out on stage.
00:21:27
Thank you. You're welcome
00:21:32
to join in and and, and take a breath.
00:21:34
This is not a moment of Zen.
00:21:35
We have work to do, so we're alright.
00:21:37
So that's agent.ai, the number
one professional network
00:21:40
for AI agents.
00:21:42
Also the only professional
network for AI agents.
00:21:46
Now, normally for Labs projects like this,
00:21:49
if I get a few hundred users
00:21:52
before the launch date, I'm happy.
00:21:55
That gives me feedback, gives me signal.
00:21:57
If I get a few thousand users
00:22:00
before launch date, I'm like,
this kid's got potential,
00:22:05
this one's gonna grow up
and be something special.
00:22:08
And I am thrilled
00:22:09
to announce agent.ai already has
00:22:13
47,000 users.
00:22:16
They grow up so quickly.
00:22:21
This makes it the second most
successful innovation project
00:22:25
we've done at HubSpot.
00:22:27
The first one being HubSpot CRM.
00:22:32
Alright? And what's equally
exciting is we have 1700 people
00:22:35
signed up for the wait
list for the agent builder,
00:22:38
and I'm approving people
literally every night.
00:22:40
So I'll be doing that in the
hotel room later tonight.
00:22:42
So sign up, it's fun.
00:22:45
Alright, and remember, you
don't have to be a coder
00:22:47
to be an agent builder.
00:22:49
You just have to be curious.
00:22:50
All it takes is a few easy clicks.
00:22:57
So we started by building
a UI, a user interface.
00:23:00
This part of it is probably a
little geekier than it needs
00:23:02
to be for a mainstream audience.
00:23:04
I'm a little geekier than I should
00:23:05
be for a mainstream audience.
00:23:06
It'll be okay. This will be quick.
00:23:08
Then we layer it on with an
00:23:09
application programming interface.
00:23:11
This allows applications
to be able to talk
00:23:14
to talk to HubSpot.
00:23:15
The next evolution of this
is what I call the AUI.
00:23:19
That's the agent user interface.
00:23:22
And the reason this is important
is this is what you need
00:23:25
for agents to be able
to talk to each other
00:23:26
and be able to collaborate.
00:23:28
And the cool thing about
Agent Builder is every agent
00:23:31
that's built on Agent Builder
automatically has an A UI.
00:23:35
Now this may not make
sense to some of you.
00:23:38
That's okay, but your
developers are gonna love it.
00:23:40
I promise. A movie reference,
00:23:42
that's a really subtle
one, but we'll let it go.
00:23:45
Okay, so now some of you're wondering
00:23:50
why is a customer platform
company like HubSpot building
00:23:55
an agent network?
00:24:01
It's my belief that the next
generation customer platform
00:24:04
needs to have three things.
00:24:07
It needs a smart CRM
that has unified store
00:24:10
of structured and unstructured data.
00:24:12
Yamini talked about this and that
00:24:13
foundational AI has to be
built into the core CRM.
00:24:17
The batteries have to be included, right?
00:24:19
Otherwise it's not a smart CRM.
00:24:21
It needs smart engagement apps.
00:24:23
You saw Andy talk about this,
00:24:24
where we're infusing AI into all the
00:24:26
engagement apps at HubSpot.
00:24:28
And finally, you need an agent ecosystem,
00:24:32
just like we have an app
ecosystem, now shout out
00:24:35
to all the app partners
that are out there.
00:24:37
Thank you. Alright, so at HubSpot we like
00:24:42
to dream big and iterate small.
00:24:46
And our dream has always
been to help millions
00:24:48
of organizations grow better.
00:24:50
We do this by demystifying, democratizing,
00:24:54
and delivering on the
potential of new technology.
00:24:58
If you're a Fortune 1000
company with a fortune
00:25:02
to spend on AI, there are lots
00:25:05
of good options out there for you.
00:25:07
But for those of you that just
want to get growing with AI,
00:25:10
HubSpot's here for you,
our focus is not just
00:25:14
to make it exceptionally
easy to use agents,
00:25:17
but exceptionally easy to
build agents and share agents
00:25:20
and distribute agents.
00:25:22
That's the idea. Now, since
our founding 18 years ago,
00:25:26
we've been inspired a lot by Apple.
00:25:31
Apple changed the world years
ago with the introduction
00:25:35
of the iPhone,
00:25:37
but what was equally
impactful was the launch
00:25:40
of the app store.
00:25:42
Their vision was embodied in the
00:25:43
phrase, there's an app for that.
00:25:47
Our vision for the future is:
there's an agent for that.
00:25:54
For every marketing, sales,
00:25:55
customer service use case imaginable,
00:25:58
including use cases we can't imagine yet.
00:26:02
There will be an agent for that.
00:26:05
Agents are the new apps,
00:26:08
and we expect someday for there
00:26:10
to be thousands of agents available.
00:26:12
We want a thousand agents to bloom.
00:26:14
Only a small fraction
of them will be written
00:26:17
and built by HubSpot.
00:26:19
Most of them will be built
by our partners, by all
00:26:22
of you, the community.
00:26:25
So we see a world where agents
00:26:27
and humans that hire them
are connected on a network
00:26:30
and can collaborate in
this vibrant marketplace.
00:26:35
Now, if we think of agents as tools,
00:26:38
we will get wonderful tools,
00:26:43
but we should skate to
where the puck is going.
00:26:48
If we think of agents as
future digital teammates,
00:26:53
we will get wondrous
00:26:56
transformation for our businesses.
00:27:00
Now, it's natural for
this to cause some anxiety
00:27:05
because it's not natural
to think of a world
00:27:08
where agents are members of our team.
00:27:11
Deep down inside, we may be worried
00:27:14
that AI will replace us as humans.
00:27:18
As a human myself, I will humbly submit
00:27:21
that AI does not reduce your value.
00:27:26
It raises it. Even
though I use AI every day
00:27:30
to help me code, help me
develop presentations,
00:27:34
help me write dad jokes.
00:27:36
I think I'm no less valuable.
00:27:38
Now I'm more valuable with AI.
00:27:41
And the same is true for all of you
00:27:43
because humans are more than the sum
00:27:46
of the tasks that they do.
00:27:50
There is an inexplicable magic in us
00:27:53
made from the experiences we've had.
00:27:56
AI may be faster than us.
00:27:58
AI may be smarter than us,
00:28:00
but it can't be more human than us.
00:28:05
Can't be because it had. It
has not had human experiences.
00:28:10
Like the thrill of starting
something new with a friend,
00:28:15
with thank you.
00:28:22
Or the profound joy of
sparking someone new
00:28:26
with the love of your life.
00:28:30
Thank you. Or the deep,
00:28:33
deep gratitude one feels
towards a new friend
00:28:37
that is helping continue
the dream and vision.
00:28:40
AI will not have those feelings
00:28:42
because AI cannot have
those human experiences.
00:28:46
It can read books, it
can inhale the entirety
00:28:49
of the internet, but that's not the same
00:28:52
as having those human experiences.
00:28:56
Having said that,
00:28:58
although AI has its limits, it's not going
00:29:01
to replace us as humans.
00:29:03
AI agents can be wonderfully helpful.
00:29:07
So we should let agents
automate the mundane
00:29:12
and amplify the magic.
00:29:14
The magic that makes you, you.
00:29:17
It is always an honor to be here.
00:29:20
Thank you so much for
coming out to INBOUND.
00:29:22
Thank you for all the support.
00:29:23
On behalf of Yemeni, Andy
00:29:25
and the global team of 8,000
plus HubSpoters, I appreciate.
![[Review] Get Good with Money: (Tiffany The Budgetnista Aliche) Summarized](https://ruhnrawpgxrzdrgaohnf.supabase.co/storage/v1/object/public/images/video/HRHktWi7AaY/thumbnail.jpg)
