The AI Workforce Engine
OpenAI is responsible for the engine fueling this AI Revolution. Learn how these models work and what SaaS leaders need to understand to fully leverage AI Workers from their Head of Strategic Marketing, Dane Vahey.
All right.
Today, I'm super pleased to be joined by an old colleague, Dane Vahiv, Head of
Strategic
Marketing for OpenAI.
Dane, awesome to see you again, and thanks so much for joining our AI workforce
summit.
Love it to see, Sean.
Good to be here and happy to join.
Awesome.
So, hey, Dane, let's jump into it.
It's pretty clear at this point that like every business leader on the planet
is just
thinking about AI is going to impact their business.
And it's pretty obvious that our investors, shareholders, and our customers are
putting
a lot of pressure on the C-suite to have like a blueprint in terms of
transformation and
how jet AI specifically is going to be embedded into the core of various
business functions
within our organization.
And marketing is really kind of one of the areas people are talking a lot about
This is obviously something near and dear to your heart, especially as it
relates to
what people are calling a marketing productivity revolution, right?
So, real high level, Dane, like what is some general advice for the marketing
leader
who's out there listening today, who are kind of tasked with having this plan
and this blueprint
to execute on this big opportunity?
I mean, the good news is marketing is the number one user we've seen, JatchyBT.
We actually look across all different types of organizations, and we do these
surveys
with our customers, and marketing is the number one of adopting the product.
So, I think marketers are actually really leading from the frontier.
Interestingly, sales is like number 10.
So even within the go-to-market organization, there's actually a lot of
difference between
kind of the marketing speed of adoption versus sales.
I think my advice to marketers is, you know, there's a ton of obvious use cases
, and I
think that is what we've seen, you know, content creation, SEO, research, data
analysis.
But really marketers really embrace this opportunity to kind of push the
envelope for innovation
when it comes to using AI and using tools like JatchyBT.
I'll give you like a few examples of things that have really kind of impressed
me.
You know, one question at Stripe we used to always ask before we launched any
product
or do any sort of, you know, formal announcement is, what do our customers
think?
And, you know, typically this involves talking to customers, and often, you
know, last minute,
we're getting on the phone with customers and having to review our content.
But I've seen actually some really innovative marketing teams use, you know,
tools like
GAU, where they can take full transcripts of concerts, of customer
conversations, and
actually create these different personas, where they can actually query that
into JatchyBT,
these different customers, these, you know, representing, you know, whole bunch
of different
spectrums, and actually get real-time feedback on products, on content, on
design as well,
and have this diversity of opinion.
So you have all these applications and all this data that is being used by a go
-to-market
organization that you can actually feed into better and form your marketing
strategy.
And often marketers are, you know, and have the best intentions to get a lot of
feedback,
but it's often done with a lot of intuition and done in kind of what's worked.
And this will really help kind of stress test marketers.
So that's kind of one advice I would say is like actually use all the data that
you have,
kind of coming into your different systems, and make sure you can try to figure
out how
to fuel that into your AI.
I think the second example that I love and like I want marketers to be
empowered here
is like as a marketer, I've never been invited to a hackathon, right?
Like we're not seen as like a technical really force within a company, but AI
is kind of
the great leveler in terms of giving anyone the organization kind of these
capabilities
where they don't need to have, you know, technical aptitude or be a software
engineer to develop
new applications, develop new products.
It really is a universal opportunity.
And so I love marketers like sponsor more hackathons, two more prompt
competitions as
well.
And a lot of the traditional blockers that you've added in your organization,
whether
it's design resources, whether it's, you know, getting opportunities to code
for a website,
a lot of those can actually be really overcome.
So we see marketers even increasing their human capacity.
So I think marketers actually be leading and kind of taking the opportunity to
really
push from the front as the number one adopters of really kind of, you know,
particularly
gents at AI tech technologies and really helping kind of push these out and be
leaders
in terms of how they should be talking about the technology.
So I think it's exciting era of kind of marketers being empowered with AI.
Yeah, absolutely.
And look, when I think about open AI, like the term that comes to mind for me
really
is accessible AI.
And there's a billboard on the one on one.
I pass all the time that says, you know, English or whenever a language you
speak is the new
sequel, right?
Like you have immediate access to all this data that you can unlock, all this
productivity
by the general end user, which is really powerful.
One of the, you know, hot topics around AI obviously is the sort of how it's
going to
impact the workforce, right?
And obviously there's lots of conversations about like, is it going to kill
jobs?
Is it going to replace jobs and all these things?
And we obviously think about it a little bit different.
I mean, I think I've heard the term or the phrase, you know, AI is not coming
for your
job, but somebody that uses AI probably will come for your job.
And we think about it in terms of, you know, not jobs, but actually like tasks
and workflows
that are ripe for AI capabilities to kind of take over and automate allowing
people to
kind of focus on different things, right?
So, Dan, you've obviously has a have a really strong history with some really
strategic roles,
including working with myself at Salesforce.
How would you characterize the impact AI will have on like marketing teams over
say the
next five years?
Yeah, I think the task orientations were a way to approach this, Sean.
We actually kind of look at it in some framework, right?
I think what you've seen today, kind of like the 3.5 miles actually, we're
really good
at replacing, you know, five second as we worked on GPT for and now with for oh
, we're
seeing kind of like these five minute tasks will be replaced as well for a lot
of marketers.
I think as we continue to evolve the kind of the reasoning and capabilities of
these
bottles, we might start seeing like five hour tasks or even five day tasks get
replaced
as well.
And kind of what the promise of this is that, you know, marketers should be
able to work
at more strategic and higher capacity type assignments and less on kind of that
day
to day.
So we really see this is an opportunity to really kind of enhance human
potential and
it really increase human capacity for achievement.
So I think like what makes me really optimistic as a marketer particularly is
that, you know,
I'll spend less time kind of thinking about document creation of, you know,
what are the
words I'm putting on a page and much more about actually like strategy impact.
I'll be able to kind of pull from a lot more information.
I'll be able to get a lot more Indians.
I'll be able to actually really kind of compliment and kind of think through my
own judgment
and my own decision making.
And then I think what's really exciting is like this individual skill
development and
growth.
I think this idea and this might be a little bit of shit from markers but like
the specialized
marketer, I think that will actually kind of maybe go into a little bit of
different
direction, right?
Because you'll have marketers that will just have such a wide range of skills
where instead
of just being an SEO marketer or someone that is really focused particularly
just on like
content creation and copywriting, you're going to get these much more well-
rounded marketers
where they're going to stand kind of throughout the whole buyer journey, how
they can drive
impact and feel really empowered to do that.
We're already seeing a lot of marketers, a lot of salespeople too, a lot of SDR
s that
can really elevate their skills and kind of run full kind of marketing
campaigns based
on using this technology.
So I think we're going to move a little bit away from a specialization.
I think people are going to have a lot more impact throughout the whole
marketing funnel.
And I think as me as someone that's hired a lot of marketers, like I love
athletes.
I love people that are willing to understand and learn from other disciplines
and empowered
by different skill sets.
I think AI is going to really help with that skill development.
And a lot of people would be much more strategic marketers.
So I'm excited to see kind of like the next breed of marketers kind of come
through this
AI generation.
Yeah, 100%.
I mean, you mentioned the SDR, right?
This is definitely one of those roles that you can see transforming overnight
in terms
of a lot of the capabilities that AI brings to that job or role or function.
Obviously, like we've shipped our AI worker, our name is Piper, specifically
designed for
that one thing.
That's what we think about.
Like what are the tasks and workflows that they execute every day and what will
open
AI provide in terms of a foundation model that will help us to automate a lot
of those
things to allow them to focus on other types of more general purpose, things
that leverage
their domain expertise, but they don't have to do the repetition.
So great, great way to think about it.
Totally agree.
And Dane, you've worked at big companies.
You've worked at startups.
And now, of course, you're on the frontier at open AI.
So I would imagine you see a huge array of different AI products, right, in
your role.
Where are you seeing teams have like immediate success in operationalizing AI
in sort of
their day-to-day marketing?
And I think if I'm correct, I think you actually even have a demo.
Yeah, I was going to show a few examples of different use cases.
How do you walk through that?
I think at a high level.
I think the teams are most successful with AI or the ones that are kind of all
ahead,
right?
You can't jump in the pool here and get like half wet, right?
You have to like, you know, really kind of embrace the technology.
We see our CEOs, like the ones that are most adamant pushing this down, it's
like their
teams should be using Chachi BT and other AI applications like 20 times a day.
And I think what I even I try to orient myself and like, I remind myself, you
know, we just
launched Chachi BT in November 2022.
It's less than two years old, right?
So we're just kind of coming and learning about this technology.
But I before I do anything, I kind of ask myself, is like, can Chachi BT do
that, right?
And I think kind of getting into that mindset of like retraining yourself,
whether it's
like doing something very basic, like putting a sticky note on your on your
laptop being
like, hey, you know, ask Chachi BT first.
But I think getting into that mindset of like using AI for every single task
will like
build that behavior.
And we really want to make that that jump where it's not like a weekly thing,
but it's,
you know, all those five minute, those five second tasks that you have lining
up, you're
using AI for that.
That'll add a lot of like value and savings in your time to make you really
impactful.
So I kind of want to walk through like examples.
I'm showing our new foro capabilities with voice and vision to like share some
ways I've
seen like sales and marketing teams start really embracing AI.
So let me let me share my screen.
You know how much I love a demo, Dan.
So I'm excited and the multi modal, you know, innovation that you guys have
happened at
OpenAI.
I'm sure is opening up a brand new world for marketers.
It certainly is.
And you know, this was, we showed this recently a couple of months ago.
This is leveraging our new kind of foro technology with voice and vision, our
advanced voice.
So this is something that's not, you know, available yet.
That's something that we're hoping to offer very soon.
It's anyone using our products in kind of a little bit of context before I go
through
our demo gear on kind of what we're showing.
Because prior to GPT for, we basically use, obviously we had voice capabilities
with
GPT for our sorry, prior for foro, we had voice capabilities with GPT for.
But when we talk to it, it wasn't very fast.
You know, there was kind of like this latency is every three or four seconds.
It was like a little better than talking to Alexa, but not much better than
talking Alexa.
And the problem is like three modalities at once, right?
You add, you know, text to basically taking voice to text, text kind of
interpreting this
and then pushing back to audio.
So you lost all that context, you lost the ability to look at background voices
, to understand,
you know, context to like understand laughter and emotion.
So we built this new model and this is for, oh, which I'm going to show you
today, which
combines all these different text, vision, audio into one modality.
So basically the inputs and outputs are processed on the same network.
So I think this is really exciting.
We're just at the start of this as well.
So, you know, always a little bit of risk showing a live demo.
But I think this is really a cool way of how markers are going to be doing this
So I'm going to jump on.
So I'm going to hit here, which is basically pulling up our voice capabilities.
Hey, chat GPT.
Hey, what's up?
Cool.
What are you up to today?
I am on a qualified call for their conference right now.
And I'd love to show him a demo.
You down to help me with that?
I'd be happy to help.
What do you need for your demo?
Well, I'm going to pull up a slide and Sean and I worked at Salesforce together
So I thought this would be a fitting slide to show.
I'm going to share my screen.
Let me know if you can see the slide.
I see the slide now.
It's showing spills force customer 360 with Slack in the center.
Great.
Is this what you're going for?
This is.
So I let's pretend I'm an AE and I got this slide, right?
And I have to go and pitch a customer on this.
I need a really great narrative to talk to this slide.
Now there's a lot of information here.
Can you help me navigate what I should say on this slide?
Absolutely.
Start by introducing Salesforce customer 360 and how it helps businesses with a
full
istic view of customer interactions.
Highlight that Slack is central, emphasizing its integration and collaboration
capabilities
across departments.
Then touch on the key segments like experience, intelligence and single source
of truth, explaining
how they enable a personalized efficient customer journey.
You got it.
I think this sounds great.
The only problem is the customer on pitching is Morgan Stanley and they're a
big financial
services firm.
How can I tailor to this pitch to be really specific for Morgan Stanley?
Good point.
Emphasize security and compliance first.
Learning how Salesforce customer 360 ensures data privacy and regulatory
compliance.
Key concerns for a financial services firm.
Highlight the platform's ability to deliver personalized customer experiences
at scale,
using data-driven insights to enhance client relationships and investment
strategies.
I like also.
But that should be T.
That sounds really good.
Are there any words I should avoid?
Any trigger words that might set Morgan Stanley off that I should not say on
this pitch?
In terms that might sound overly ambitious or promise dramatic results like
disruptive
or game changer.
Also steer clear of emphasizing rapid change or risk taking as financial
services firms
are generally cautious.
Instead, focus on stability, reliability and how Salesforce can help them
innovate
while managing risk.
Does that cover your concerns?
I think you nailed it.
All right.
Thanks for your help chat, Jimmy.
You're welcome.
Good luck with the pitch.
Reach out anytime you need more help.
That's one example of how I might use chat GBT on a sales case.
I actually want to show a marketing example too.
For those of you that don't live in the demand-gen world, something like this
might be a little
intimidating.
You get this report.
It's looking at a bunch of different data.
Different channels here.
You have different metrics, cost per lead, leads closing and so forth.
If you get this, you want to be able to ask the right questions to know how to
interpret
this strategy.
I think what you look powerful about vision is it can look at data like this
and really
help synthesize what you should be extrapolating from information like this.
Let me have chat GBT help me look at this data again.
Hey, chat GBT.
Can you see all the lists of data and tables and charts I have here on my
screen?
Yes, I see the data and charts.
There are various graphs showing leads, costs and other metrics over time.
What would you like to analyze or discuss about this data?
Yes.
I have to understand this data on meeting with our demand-gen leader.
Can you look at these different channels between email, SEO, content and tell
me what channels
are actually being the most effective for our marketing team?
Looking at the charts, the first three graphs show leads generated, cost per
lead CPL and
percentage of leads closed by channel.
Email seems very effective, consistently generating a high number of leads with
a relatively low
CPL and a good percentage of leads closed.
SEO also performs well in terms of leads generated and CPL, but the closure
rate is lower compared
to email.
Content marketing has fluctuating results.
A great all-men.
A trashed chat GBT.
So it sounds like email is a great channel.
So I would love to come in with some proactive ideas for how we can really take
email to the
next level.
Looking at this data, what would you suggest we do to really improve our email
performance
even further?
Based on the data, I'd suggest focusing on optimizing your email content and
targeting.
Try A/B testing different subject lines, content formats and calls to action to
see what resonates
best with your audience.
You could also segment your email lists more precisely based on customer
behavior or demographics
to deliver more personalized content.
Also, consider automating follow-up emails based on user interactions to
nurture leads
more effectively.
Does that help spark some ideas?
That does.
Alright, I think I know what I need to do for this meeting.
I appreciate your help chat GBT.
You're welcome.
I'm glad I could help.
If you need more assistance, feel free to reach out.
Take care.
So, let me stop sharing my screen here.
Hopefully, that was a helpful demo of just kind of walking through a few of the
vision
capabilities.
You can see as this becomes very easy to use, pulling up your screen, talking
to your computer,
pulling up data, pulling up different slides, understanding contacts.
I really kind of think markers will be using this every single day and just be
able to
kind of the convenience of the use cases.
I know it's something that I love using.
Dan, thanks.
I really appreciate that.
A couple of things.
You can immediately see the workforce impact from that demo.
You can see sales enablement, time saved.
You can see product marketing, time saved.
You can see marketing ops or business analyst, time saved.
It's so amazing that you can have this really natural language interface
allowing you to
get work done.
The other thing I liked, I love the fact that you get interrupt chat GBT when
it starts
getting verbose and just get onto the next thing.
Super fascinating demo.
Every time I see a demo from you guys, it just gets you excited about being in
the software
industry.
It's as exciting as it's ever been possibly much more exciting.
It's really funny you say that because talking to my Alexa, it's a rising crazy
doubt.
I just want to interrupt.
Who would have known the power feature for human communication was like
interruption?
Maybe we should all be better listeners.
You'll be interrupted when it's talking and clarify and so forth into your
contacts is
by far the best thing is about the set.
Really thoughtful because we know that LLMs can be verbose.
That's an amazing demo.
I really appreciate how applied that was to you can imagine your job over time.
I heard a statistic that said in marketing, people are going to be 50% more
productive
within the next 12 months, which is extraordinary.
Think about even 40%.
That's like a day and a half off a week, not that they'll ever let you take
that off,
but impressive.
I think that's why people are talking about this productivity revolution as it
relates
to AI and especially as it pertains to marketing.
Last question for you, Dane.
Let's go future a little bit.
You guys are obviously on the front lines of AI development.
The foundation model is really disrupting the way that people work in general
and quite
honestly the way people live.
What do you think we need to be prepared for in the near future in terms of the
next revolution
in the AI era?
Thinking ahead a little bit, what should be preparing for?
What excites me is this idea of hyper-personalized content.
We started this already with memory in chat GPT where it knows information
about me and
I think what was powerful is I can, unlike real life, I can selectively edit my
memory.
I could say what I want to remember and what I don't.
As you build this more and it becomes more capable and you start using these
models.
Again, chat GPT is less than two years old, but imagine you using chat GPT and
you have
this history of eight, nine years of data, of interactions, of experience.
It's going to understand your preferences, your style, your past history really
well.
You could enter in all the documents I've ever created.
Eventually, I can take all my emails and put them in there, or really my tone
of voice.
I think what we're just going to see right now is this basic version of chat G
PT in the
sense that it's feeding on and you can customize it obviously in a lot of ways.
It's not feeding how you interpret and how you interact with it.
I think we're just going to see a lot more sophistication, a lot more personal
ization.
Through that data, you're going to start seeing, you already have a lot of
companies doing
this today with our assistance API of building this agent behavior where AI can
help automate
a lot of these tasks on your behalf.
I think that's a really powerful frontier.
Imagine Sean, finding time on the schedule.
We had two agents that were able to communicate with our calendars, find time
collectively,
know the things that work well for you and I, schedule is meeting.
That's a powerful time saver, then going back and forth and having to interact
and find
opportunities.
I think the idea of this hyper personalization, really guided self-learning
development,
this really nuanced actions and these agents that help you do all these tasks
in automation.
I think a 50% of the activity might be low.
I think that, for you have a lot more time, it's actually working on a lot more
strategic
and powerful work and take a lot more on.
I'm really excited to see the savings of all these automated road tasks and the
personalization
level that's coming to AI.
Yeah, no, that's a great point.
People forget that we're in the first step of a marathon and that, you know,
chat GPT
hasn't been around all that long, right?
I think I saw a statistic from, I think it was Alex Wang on the 20 VC pod
talking about
GPT-3 was trained on less than a petabyte of data, trained on like the internet
You imagine the vast amount of frontier data and then you start thinking about
historical
data and how that's applied.
You can think that personalization is going to be extraordinary.
So, hey, Dane, really appreciate you joining us today and your insights and
really love
the demo.
That was really great and it was good to connect with you again and I hope we
get to connect
in person soon.
Yeah, wonderful.
And if anyone wants to join me for a marketing led hackathon, let me know,
sustain at OpenAI.
We'd love to speak to more people building an AI.
I think the other thing I love about my job is every time I chat with a
customer, I hear
kind of a wonderful or brand new use case.
There's really no box kind of limiting the potential for what people are doing.
So, great things you're doing.
I love to hear about them as well.
So feel free to reach out and thanks for time, Sean.
I really appreciate joining you and everyone else here in the qualified world
for this talk.
Awesome.
Thanks, Dane.