Subscribe
Share
Share
Embed
A podcast that explores how AI is transforming careers, businesses, and industries. Hosts Greg Boone and Erica Rooney deliver real-world use cases and actionable AI strategies to help professionals stay ahead of the curve.
Jess Lipson: I would say one of the keys
is getting good written documentation of
everything that you do in your business.
So I think like one of the differences
in, uh, training in AI versus.
A human employee is that a lot of human
employees can kinda learn by osmosis.
So you don't necessarily have this awesome
onboarding doc with all the background
of everything they need to know.
They just kind of come in, they meet
with people, they ask questions,
and over time they figure it out.
Greg Boone: AI isn't the future,
it's now, and whether you're in hr,
sales, operations, or leadership,
the choices you make today.
We'll determine whether you
thrive or get left Behind.
Erica Rooney: Today's guest is
Jesse Lipson, founder of ShareFile,
and now Cee O of Levitate, where
he's helping over 7,000 businesses
scale client relationships with
AI while keeping it personal.
Jesse, welcome to the AI Voicer victim.
How's it going today?
It's going great.
Amazing.
Well, I really wanna kick it off
and start talking about how you have
scaled several companies from scratch.
I wanna talk about your thinking around
innovation and technology and how has that
evolved since your early share file days?
Jess Lipson: Well, my early
share file days, um, really kind
of the beginning of the cloud.
So I think the, kind of, the thing
that where we got right or lucky was
I launched ShareFile at late 2005 and.
Term, like if you look at Google Trends,
the term cloud didn't even, um, happen
for a couple more years after that.
And so, um, in ShareFile it was like
kind of, there was so much opportunity,
it was just the beginning of software
as a service and web-based apps and we
were able to kind of ride that trend
and there was so much white space out
there in, in software and in tech.
So I think how it's changed, I mean it's
been 20 years since then and I. Um, the
marketplace is a lot more sophisticated.
I remember with share file, uh, Google
AdWords had just come out and ads were so
cheap and so you could just, any software
could advertise and for, you know, 10
cents a click be getting great traffic.
And so like, yeah, it's
a lot more competitive.
Um, the world of software these
days, and I think software.
One of the things I realized when I
started Levitate is that software,
you're not just competing with, uh,
other products in your category.
Business users only have so many
products they're gonna use, so you're
basically competing for attention with.
Every other software that's out
there, um, kind of these days.
And so it's just, it's a lot more
sophisticated, a lot more competitive,
and a lot more difficult to get users'
attention now than it was 20 years ago.
Erica Rooney: Well, I'm here for
some 10 cents click, but those,
those are some days of the past.
But tell me a little bit more
about levitate and what you do
and how it really incorporates ai.
Jess Lipson: Yeah, so Levitate is
basically, um, a marketing software for
micro small businesses, so companies
that are about one to 25 employees.
We specialize in certain industries
that are, tend to be more relationship
based referral, word of mouth,
kind of driven, like insurance,
accounting, legal, nonprofit.
And basically what we help them do is.
Keep in touch with their contacts in
a much more personal, authentic way
at scale than a traditional marketing
software that's more of like a mass
blast kind of newsletter type approach.
So kinda one way to think about it is.
Let's say you are a, um, wealth
manager and you have a hundred
clients or a couple hundred clients.
We help you keep in touch with those
clients consistently in a, in a
personal, authentic way versus just
sending a HT ML newsletter, which most
people these days don't really read.
And even if they do, it's not, it
doesn't deepen the relationship in the
same way that a personal outreach does.
Erica Rooney: Oh my gosh, no.
We are in the age of hyper-personalization
and where that is so, so key, especially.
With AI being everywhere and people
just getting inundated with the
same types of content being seen.
Yeah.
So I imagine that is extremely
successful and is gonna be even
more necessary in the coming days as
more and more people lean into ai.
Jess Lipson: Yeah, so when I
started Levitate, it was the
end of 2017 and we were actually
called levitate.ai back then.
So my thought was how
can we use AI to help?
Um, help you scale personal, authentic
outreach by identifying contacts that
you want to be keeping in touch with,
but have slipped through the cracks.
And so the original idea was we connect
to your email and calendar system, which
we still do, and we can look at different
patterns and say, Hey, like, you want to
be in touch with Greg, but you haven't.
And, and, um, like I think all of us
in business have people that we know
we should be keeping in touch with.
It would help us if we did, but.
Life gets in the way.
You get busy.
And so that was the problem
we were trying to solve.
But for probably the first five years,
the fact that we were called Levitate
AI was kind of a running joke internally
because we just didn't use very much ai.
That was part of the vision, but
the technology wasn't there yet.
And now over the last couple years
we've been able to layer it in and
and probably really start to achieve.
Erica Rooney: So you dropped the
ai, but now we could actually use,
Jess Lipson: oh, we actually
kept it the whole time.
Luckily we started with the DO ai.
Yeah, we started with it, you know,
it didn't mean much for a long
time, and then now it's y'all.
That's a
Erica Rooney: manifestation at its finest.
Oh my gosh.
Okay.
So you've obviously been on the front
lines of AI since the beginning.
Tell me, what does responsible
practical AI integration look
like for small businesses?
Right now in today's landscape,
Jess Lipson: I think it depends
on the type of small business.
Uh, it's very different if you're a
company like us that's a software company.
But I think that for, um, your traditional
mom and pop small business that the
types of small businesses that we
serve, it's still very, very early.
Um, and I don't see a lot
of small businesses yet.
That are adopting AI in a major
way, kind of at their core.
I think a lot of people are still
trying to figure out how to do it.
I mean, the common use case a lot of small
businesses use is just like what average
consumers use, which is just, um, using
chat GPT to help with a lot of stuff
that they would normally use Google with,
with, or ask, you know, somebody about.
So just like, Hey, I need to, um.
Write this business plan, can you help me?
Or I've got, I wanna draft this
article, can chat GBT, help me edit it.
That kind of thing.
Or I've got an email, a really
critical email I need to write where
I'm really mad about something.
I push it into chat GBT and
ask them to help me make it
nicer or something like that.
Speaker 4: Erica used that one a lot.
Jess Lipson: Exactly.
So I think there's kind
of around the edges.
Um, one of the challenges for small
businesses and, you know, we can
get into this a little bit more,
um, in the conversation, 'cause I
do think it's important is that.
I think to get AI into a core process
that is, say like, um, replacing some
of your customer support burden or sales
burden, uh, it actually requires a decent
amount of investment for kind of training.
Um, and so I think that, um, a lot of
businesses are not, have not yet crossed
into that investment of like really
trying to get it to displace support.
Queries and things like that.
I think some, actually, some larger
enterprises are, um, a little bit
further along there because they
have the ability to make larger
investments in kind of tuning something.
Erica Rooney: Do you think though,
that smaller businesses should take
more time because they don't have that
type of capital to wait and see what
happens with all of these different
technologies that are out there?
Because.
I mean, obviously AI is the
hottest thing out there right now.
Everybody's doing ai.
There's a million different
technologies, kinda like HR software.
There's 15,000 payroll systems you can do.
Do you think it's better for them
to wait a little bit until it kind
of dies down and the leaders in
the marketplace rise up to the top?
The people you know who can really
produce the best that way they don't
spend, you know, small dollars and get
Jess Lipson: small potatoes.
I think there's kinda layers.
So like the first layer that's easy and
everyone should do is, you know, use.
Chat gt a lot, please.
And especially, you know, um, some
of the getting a, what I would
recommend for a small business is
to get a teams account because, um,
that can also protect your data.
So you can control chat, gt
not training on your data.
Um, and uh, and so I would say like.
Get a team account for your business,
get your employees in the habit of
using it to help with productivity.
And that probably, you know, gets you a
little productivity bump just doing that.
I think the second layer, um, which
is also if you do have a team account
for Jet GPT or a paid account is um,
what they call custom GPTs, where
you don't have to be a programmer.
That's where you can start to
dip your toe in a little bit of
training and customizing to make
it more useful for your business.
Um.
So you could have, hey, a custom GBT
where you upload some documents of
questions that you know clients would
normally ask you, and you give that to
your employees and they can use that.
And that's kind of one level above what
a, just a untrained chat GBT would be.
And then I think the third
layer is, can you use it to, um.
Help answer sales questions and support
questions and or integrate it into things
like spreadsheet work that you're doing.
And I think that that takes the next
level of commitment and investment.
I think the technology and tools are
there, but, um, it's more about m
making the commitment be because you're
not just gonna be able to, you're
not, probably not gonna have a great.
Um, AI customer support agent.
If all you do is just.
Take the vanilla thing and maybe
point it at your support documents.
You need to do a lot more, you know,
reinforcement and customization in
order to make it truly valuable.
Erica Rooney: Yeah.
I'm a big believer in getting
people to really understand how
to create those custom chat.
Chat.
Mm-hmm.
Chat bots.
Yeah.
Because I think number one, as someone who
considers herself very non-techie mm-hmm.
I had a lot of fear around that at first.
I'm not techie.
I dunno how to do that.
But when I saw how easy
it was to do that, yeah.
I was like, holy shit, more
people need to be doing this.
Yep.
And now I have a whole army of them and
I go out and I try to make sure that if
anybody does anything, which one is your
Speaker 4: favorite?
Erica?
Oh,
Erica Rooney: I got right now.
Well, I got Difference.
The one right now that's my
favorite is my newsletter writer.
'cause she's been cranking out some,
Speaker 4: what's her name?
Erica Rooney: Nancy,
the newsletter writer,
Speaker 4: Nancy, she likes to name them.
They're
Erica Rooney: all females
Speaker 4: like Cheryl.
Cheryl is her, uh,
Erica Rooney: Cheryl's my
twin after Cheryl Sandberg.
She does a lot.
Yep.
And then we've got Nancy, the newsletter
writer, and Caroline the content queen.
I've got a lot of different ones.
Yeah.
Right.
But I think the, the reason that I
love that you say please just start
leaning into these teams, start getting
a little bit of productivity wins, is
because it's a great primer for change.
Yep.
I think had I not.
Taken the chance on myself to
create that own chat bot, I
probably still would've been very
like, I don't know about all this.
And so leaning into that is huge.
What would be the next step do you
think, for businesses, if they can
get that team's account, what's that
next critical step for them to help
with transforming their workplace?
I
Jess Lipson: would say one of the keys
is getting good written documentation of
everything that you do in your business.
So I think like one of the differences in.
Training in AI versus a human
employee is that a lot of human
employees can kinda learn by osmosis.
So you don't necessarily have this awesome
onboarding doc with all the background
of everything they need to know.
They just kind of come in, they meet
with people, they ask questions,
and over time they figure it out.
Um, and so a great way to be prepared
for AI is like, let's say, um, take the.
The customer support case or the
newsletter case, or let's say a
case with sales, it's like, okay,
well how would we train SDR?
What are the principles of.
The five steps of, you know, a sales
process and then what are common
objections that people have and how
would we address those objections?
And you have to do the pretty hard work
of actually documenting and writing
all that stuff down, which is actually
something that you probably should
be doing for your employees anyway.
But most companies haven't done
that, especially like small
businesses, but even large businesses.
And so that will benefit
your employees anyway.
But I think having that.
Is a, um, prerequisite to doing a
really good job getting a GPT, you
know, trained for your business.
Speaker 4: Yeah, we spend
a lot of time, um, Peter.
First of all, thank you
again for joining us.
In full disclosure, I'm
an investor in Levitate.
Um, I met Jess, I guess it's been
five or six years now through,
it's kind of, kind of roundabout.
I actually was introduced to
Howard Lehrman who then introduced
me to you because Howard said,
Hey, you need to meet Jess.
You guys are local.
I was like, well, I know
Jesse's wife Brooks.
'cause I, you know, we
competed against each other.
We partnered on things in
the, in the past and, uh.
I remember in the early days, I
asked Jess, I said, Hey man, how do
I, how do I get in on this AI thing?
I don't know what this is.
Yeah.
In like 20 19, 20 20.
And then a few years later
I was able to, to get in.
But you know, you said a lot
of things here that, uh, I
just wanted to just kind of.
Wind back a little bit.
Sure.
If you don't mind, like, one of the things
I think it's very imperative for folks to
understand is, even here at Walk West, we
talk about making, you know, AI adoption.
Three things, make it personal,
make it fun, make it safe, right?
We have a, our own competition
that we're, we're doing right now.
'cause we do believe that we need
to create these unlocks, right?
These, uh.
I've trademarked AI serious.
She calls herself AI Curious.
We talk about moving people from
anxious to curious to serious.
Yeah.
And the idea being how do we get
people more curious internally, right.
And then get to that point where, as
you described, documenting the process,
I say, Hey, look, I always take a
people process platform in that order.
Always approach.
Yeah.
To any consulting I do.
So right now we're focused on the
people, but we have to go take
a look at the process, right?
This is not a situation
where we just bolt on ai.
To existing process.
This is an opportunity to reimagine
what the process, what the team, you
know, I hate to use the word reorg,
but structure it in a different way
than what you've done in the past.
So do you find that the smaller
the organization, the easier to
do that or harder to do that?
Jess Lipson: I think it, um,
there's some things that are
easier and some things are harder.
I think a small organization.
Can change and move faster, but also has
less resources and a larger organization,
you know, everything in a large
enterprise seems to take multiple years.
Um, but.
There could, they potentially could have
the resources to say, we're gonna put
five people on this, or we're gonna do
a lot of investment on the engineering
side to, to make something happen.
I think for technology companies like
software, um, there's a big advantage
for companies that have, are AI
native that have basically started
in the last couple years because, um.
From the ground floor, they're
using AI tools for building
their software and adapting.
And I think, um, for established
companies, there's, and people that
have a lot of experience, I think
there's always a version to change.
And like in every department, um,
getting a sales team to use AI tools
takes a lot of change management.
And there's gonna be resistance where
if you, you know, if you just start.
In the world of sales in 20 24, 20 25,
you're just gonna naturally embrace some
of those tools, um, a lot more easily.
So I think there's challenges.
Uh, I would say just in general
for a software company, a
tech company, it's probably an
advantage to be smaller right now.
Um, just in terms of moving quickly for
a non-tech company, I think it's like.
Kind of 50 50.
There's advantage to disadvantages.
You
Speaker 4: primarily work with non-tech?
We do companies right now.
Mm-hmm.
I was actually, uh, last night, um,
I was telling a buddy of mine who's
actually my chiropractor, I said,
Hey, I got a guest on tomorrow.
I said, you should probably
try out his, his technology.
You're the perfect fit.
Yeah.
You have like a five to 10 person group.
Right.
He has an identical twin brother
who also has a Cairo practice.
Oh, nice.
I was like, you guys are perfect for this.
So I'm trying to get him to come in here.
You may see him come in.
Jess Lipson: We're actually
working on, we're working on HIPAA
compliance right now, so the next
couple months we should have it.
But, uh, Cairo Dental, those
are great use cases for us.
Speaker 4: I mean, Eric and I have
historically worked in enterprise, as
you know, as far as our client base.
Yeah.
And I was telling folks, I was like, I'm
excited about this age of AI now, because
even, you know, I run Walk West now.
Mm-hmm.
Right.
I'm able to actually provide for
smaller organizations because
the, because of the tools.
Mm-hmm.
Right.
It was cost prohibitive for
us to really work downstream.
Yeah.
With even, you had to be upper middle
market to enterprise for us to really
see any value that we could provide.
Right.
But now with these tools, right.
Able to come further downstream,
was that something that you kind
of witnessed upfront or was it
more of a passion for you to work
with these smaller organizations?
Kind
Jess Lipson: of a passion.
We did share file kind of
pre-acquisition from Citrix.
We worked with small businesses just
about exclusively, and so it's kind of
what I know is the micro businesses.
I did work at Citrix with some large,
we had an enterprise part of the
business that we developed there.
But, um, I've always liked
working with small businesses.
I'm co-founder of Raleigh founded.
Um, I like working with entrepreneurs
and I think, um, the thing that
gratifies me about working with
small businesses is that they can
benefit so much from technology.
Um.
Very few software companies will try
to sell to small businesses because
they don't have a lot of money.
They're not that tech savvy.
They need a lot of support, um,
which is not usually a great
combination for a, a, a vendor to
service those types of businesses.
So, but I think the cool thing is
you can actually help 'em implement
something and it dramatically changes
their business and they really appreciate
it, whereas a large enterprise, um.
You, maybe you move their earnings
per share by a penny or two and like.
Nothing against that, but for me it's just
more gratifying if there's like, oh, this
chiropractor and they implement our stuff
and they're like, wow, this is amazing.
I get like all these clients that
used to fall off, now they're coming
back 'cause I've figured out a way
to reengage them and like my business
is up and my customers are happier.
And that just feels really good to me.
Speaker 4: That's what I was trying to
explain, you know, uh, to my Carol friend.
I said the challenge that
you guys have is that, um.
You know, a different example is I
have a dentist appointment tomorrow.
Yeah, right.
I get my teeth clean every six months.
Yeah.
There's no connection.
With that business over
that six month period.
Right.
Speaker 5: And
Speaker 4: so you have
no connective tissue.
So if they needed something in that
moment, you're not top of mind for them.
Right.
Right.
And I said, that's the
fundamental challenge.
Right.
And I wasn't even talking about levitate
Speaker 5: Yeah.
At
Speaker 4: the time.
But then I, it started to dawn me.
I was like, okay, here's the solution.
Yeah.
Right.
So I'll turn it back to Eric Oak.
Erica Rooney: No, I mean, these
are great questions and it actually
leads me to what I wanna talk about
next, which is all about enhance,
enhancing empathy and trust.
Because I know one of the biggest fears
that people have is this removal of the
connection and the human element of it.
And with levitate, it really emphasizes
that relationship driven growth.
How does one design AI tools that
maintains that level of empathy and trust?
Jess Lipson: Well, from the
beginning for us with levitate.
The goal was to, um, help people
scale their authentic outreach
versus replacing it with, you know,
completely automated approach.
'cause I think, um, I think
people have gotten really
savvy about automated outreach.
Um, I mean, before AI you had like.
20 years ago, if you sent A-H-J-M-L
newsletter, it worked great and everyone
was like, wow, you know, like I got
the newsletter from whatever company.
And um, they would click through and look
at it and then eventually they're kinda
like, uh, newsletter, um, that's not too.
Like personal, but then this whole
technology of kind of email sequences
came out of like drip campaigns.
And that kind of fooled people for
a while and it worked pretty well.
And then eventually people are like, oh,
this is, like, this is a drip campaign.
Um, and so, uh, I think people, and, and
even now people can kind of sense an AI
email and like AI content a little bit.
And so, um, what our goal was, was.
Actually we started more with the
first piece, which is, um, who should
you be reaching out to and helping,
uh, use AI to be like, look at a d
some different signals and say like,
here's 10 people that you should be
reaching out to that you haven't like.
Um, and that's also if, if I had a
meeting with Greg yesterday, then it
would be really weird if he got, you
know, an email from me tomorrow, you know?
And so I think being thoughtful about, um.
Flagging who you should reach out to.
Whereas I think most people, traditionally
it's just like I happen to run into
somebody or think about somebody and maybe
I reach out a little bit, but like 90% of
the people that I would probably like to
be keeping in touch with, I just don't.
Um, and then the second thing is we,
we try to kind of mass personalization
where we've got a workflow where.
If you do send out to a client list,
um, you may start with a template, but
you can kind of click into each person
and spend 30 seconds and add a sentence.
Um, and then the other thing is like,
this is things that we're just getting
into now that AI has gotten to where,
um, where it is, and I wanted to do at
the beginning, but we weren't able to, is
things like coming out of this meeting.
Um, I can potentially.
Just record some quick notes.
And AI is smart enough to put
those on your profile in a way,
with tags and other things.
If I find out like, um.
Do you have certain interests or hobbies
or food you like or something like that?
Teams you like?
Will
Erica Rooney: it tag up?
Like let's say you and I have
a conversation and you're like,
Hey, I'm gonna do this big 99
mile bike ride in six months.
Yeah.
Will it tag that?
So then in like five months I
can be like, Hey, how's it going?
You've got that thing in a month.
Jess Lipson: That's the vision.
And I think now with
ai, we're getting there.
And then one thing that we're about
to roll out is, um, just being able
to ask questions about your contacts
would be like, oh yeah, like, um.
Erica, like, what are
her kids' names again?
And being able to have a conversation
with your system instead of, it's
just so tedious to record the things
and look them up that most people,
especially small businesses, would never
go to that level in A-A-C-R-M system.
But I think with AI
you'll be able to be like.
Walking out and be like, all right,
um, here's a transcript from the
meeting or notes from the meeting.
And it'll pull out some tags of
like, oh yeah, like some key facts.
Like, Greg paid, played
basketball at Howard and you
know, a couple other things.
And then, um, I probably would've not.
I would either have to remember
that or which I try, but, you
know, hundreds of people and
sometimes, you know, things slip or.
I'm never gonna take the time to like
log into Salesforce and like write these
copious notes about every time I meet.
And so I think it can, uh, I think we
all have these people that are just,
we know that maybe a couple people
who are super thoughtful and they're
like, oh yeah, how is that vacation
you took like a year ago to Bora
Bora that you were telling me about?
And like, how's this and how's that?
And um, I think what we're trying
to do with technology is like.
Let the average person
kind of be extraordinary.
That thoughtful.
Yeah, exactly.
Yeah.
Erica Rooney: No, I love that.
And actually I've, I've coached many
women on how to network better, and
a lot of it is, is kind of like this,
but it's very mundane in a Google
sheet where it's like Wednesday.
Name.
When's the last time
you reached out to them?
What are some key things
you need to remember?
Yeah, like their kids' names and
it is very tedious and most people
don't keep up with it long term.
They'll use it as a short term fix, like
when they're job hunting or something.
Yeah.
But that's incredible.
I love that.
But sales,
Speaker 4: you know, to
your point about SDR, so.
Nobody wants to put all
that information in the CRS.
Yeah.
Yeah.
Right.
It's one of the, the biggest pain points.
And when, you know, when, when chat
GPT launched in, in November of 22,
you just saw people racing to go, go to
market solutions for that very reason
because the pain is so real that in
the sentiment analysis and being able
to use natural language processing to
actually have a conversation mm-hmm.
With the information to your point.
Yeah.
If I'm on my way somewhere, I'm
like, oh shit, what did Jess
say that he used to focus on?
Or this, this and this.
Yeah.
And give you, you know, we use like, um,
I know a lot of folks here use granola to
transcribe things and be able, but that
whole kind of workflow, it's different
for folks and I've seen a lot of sales
folks really gravitate towards it.
Um, I think what I like a lot
about levitate though, it's not,
this is my understanding of it.
It's not so much a point solution
like some of these other things.
It's more of a kind of a unified, um.
A solution.
Almost like a, like a mini
kind of HubSpot or something.
Yeah.
C rm.
Right?
Yeah.
And I had a chance, uh, last month to
go to, uh, HubSpot's ai, uh, summit.
It's like invitation only.
So it was good to see like faces
of like all of the founders.
Mm-hmm.
You know, and startup, uh, aspiring
startup folks was only like 300 folks.
Mm-hmm.
And then Crosstown, it was the Databricks
Conference, which is like 20,000 people.
Right.
And just kind of seeing that.
That dynamic is like, all right.
Um, what I don't want to get to,
and part of the reason why we call
this AI voice of victim is that we
don't end up at a point where there's
the haves and the have nots, right?
So I love having a guest on here
that's talking about small business.
I would say like 80, 90% of our guests are
primarily focused on enterprise mm-hmm.
Type of client.
So it is exciting to see that, like
what other barriers are you seeing?
I always tell folks, I'm like,
it's hard to sell something that
people don't know how to consume.
Yeah, right.
So even OpenAI is going through this
right now where for, what is it?
Accounts that are 10
plus million for them.
They're providing consulting services.
Yeah.
Right.
Because you can only sell so much
to your point, if smaller businesses
don't have the technical resources or
understanding, are you having to onboard
them, uh, in any significant way or
what does that look like to get them,
Jess Lipson: you know,
ready to use levitate?
So, um, we have.
We probably provide the level of
service when it comes to customer
success that, uh, of somebody who
would normally spend like a hundred
thousand dollars a year or more.
We, our customer success team meets
quarterly with each client and helps
them map out a content calendar
and strategy and things like that.
And the main reason for that is just
that small businesses need that.
If you don't provide it, um, they're.
Not gonna be active with the tool
because they have so much going on.
So I think that the key to, um, being
successful in the market with small
businesses, figuring out how to do
that and be efficient enough and
AI can be helpful behind the scenes
and helping us do that, be efficient
enough that the margins are good
enough and it's profitable, but you
really have to run it like a well-oiled
machine in order to make that work.
Erica Rooney: What AI tech
stack are you using right now?
Jess Lipson: We use open ai.
Um, that's, uh, that's really what we use.
We, um, both from a, um, chat
GBT teams, but also their API.
Um, then we also use, um, Claude
from an engineering perspective.
We use a lot of cloud code and we can,
you know, get into that, but that's
been enormously transformational.
But, um, so anthropic, uh.
Seems like they're way ahead when
it comes to the, the coding models.
Erica Rooney: No, I love to ask that
question because there's obviously
so many different tech stacks and
combinations out there that you can use
and you know, so many people have, so.
So many different options
to choose from that.
I like to just say like, it's
your own little mini bar there.
Like pick your poison, do what you
wanna do, test them out, explore
and settle on one that feels right
to you, and then spend the time to
really get to learn it and know it and
integrate into it because you might.
Surprise yourself.
Speaker 4: Mm-hmm.
I think that the, the challenge
though for a lot of organizations
is there's just so many Yeah.
Tools.
Speaker 5: Yeah.
Speaker 4: Right.
So, so many like, I was gonna ask
the question, so I'm glad that Erica
Speaker 5: Yeah.
Speaker 4: Asked you the question.
'cause I was trying to understand.
It was like, all right,
what's it underneath it?
You have your own, you
know, large English model.
You have your, your you small, like
you're, it sounds like you're, you
know, fine tuning it and using some
reinforcement learning and some
things to, to really get there.
Our, uh.
Our SVP, uh, here, uh, John
Za, a marketing technology.
He was at Hello Alice.
Most recently helping small businesses.
They were using Claude.
Yeah, right.
For a lot of the
development things, right.
To be able to get features out.
So it's good to kind of see
these parallels and folks really
trying to dive in and help at
scale these smaller organiz.
I think one of the challenges is
that I've seen is there, because
there's so many tools though,
there is this analysis paralysis.
Yeah.
Right.
Because folks are like, you know,
for us, we're a bit of a in a bubble.
Always try to explain to folks I'm like.
You know, whether you're in the
triangle or if you go to San
Francisco or some major cities.
Yeah.
It seems like a lot of people
or everyone's using it.
Yeah, but they're not.
If you go to Middle America, if you
go to many of the main, you know,
main Street type of businesses,
they're not using our understanding.
So coming to them with a stack.
Speaker 5: Mm-hmm.
Speaker 4: Right.
And saying, Hey.
Why don't you try these out?
Because there's gonna be a
thousand options in two years
for every, well, I don't know.
You're the expert on this.
I don't know when the bubble
bursts, but there's gonna be
some non winners for sure, Ray.
Yeah.
Is that what
Erica Rooney: we're the
word we're using now?
I was told
Speaker 4: that when I'm on a panel
Erica Rooney: non winners,
not to say losers anymore,
Speaker 4: like it wasn't specific
to me, but non making a note.
Non non winners.
Right.
But the the reality though, is that.
Because the barrier to cre, my
degree is in computer science,
but I, so I vibe code Yeah.
Quite a bit.
Yeah.
Now, but I can actually understand and
trace the code and like understand,
like I was literally doing something
before this, and I was telling John
and team and Tom, I said, I think
I just created an infinite loop.
Speaker 5: Uh,
Speaker 4: I, I'll see what's
happening when I come back.
Right.
I kept trying to stop it.
Yeah.
Yeah.
But it just kept going back.
So,
Jess Lipson: but yeah, no, it's, uh, I, I
mean, to your, to your question, I think
what I would say to people, and I've said
this before, just to keep it simple, is.
Just use open ai, like, you know, yeah,
there's Gemini, there's, there's a lot of
things, but just constantly poking around
every model, you know, is confusing.
Like open AI is good.
And then I would probably say if you're,
if you're coding, just use quad code.
It's the best.
Keep it that simple.
Yes, I know there's like deep
seek and gr and you know, um.
All these other tools, but, and Lama,
but I think people can get really
confused, like OpenAI is the market
leader and their models are really
good and just like keep it simple.
Speaker 4: Yeah, I think the, uh,
I mean it's, it's great advice.
We do use a lot of Gemini here
because we're a Google workspace.
Like there's a reason why, you know, but.
If we had specific engineering
related things you like, I
have open AI keys, right?
Mm-hmm.
We use those things.
We'll use Claude for other things, right?
And as being a marketing agency, we're
actually, what we're trying to do is
say, look, what are the three to five
tools or things you need to solve?
We'll give you some recommendations,
some ideas play around with these
based on your specific role, right?
If you're a content
writer or strategist, use.
Claude.
I just fundamentally
believe it writes better.
Yeah.
Right.
Mm-hmm.
You know, I think, but, uh, but
giving folks some of that, that
freedom to play around with.
To your point, um, I was listening
to a, uh, legal podcast coming in
here, and they were talking about
getting a teams or enterprise account.
You know, so that you are
not worrying about mm-hmm.
All these other things.
Like, okay, how do I
redact this or remove that?
And it is like, if you do that,
you're gonna increase such a barrier.
So much friction, no one's gonna adopt.
Speaker 5: Yeah.
Speaker 4: Right.
And so I think you've done an
amazing job, you know, again, with
your tool, it reminds me when you
were talking about screens earlier,
I wanted to chime in 'cause I was
recently listening to the 20 v.
20 VC podcast.
Speaker 5: Yeah.
And
Speaker 4: they had the chief
product officer from Duolingo.
Speaker 5: Mm-hmm.
Speaker 4: And they kept trying to
say, what category were they in?
It was like, well, are you in education?
Like we're not in education,
we're competing for screen time.
Speaker 5: Mm-hmm.
They
Speaker 4: said the same
thing that you were saying.
Yeah.
It's like we're actually
competing for screen time.
Mm-hmm.
Right.
And so whatever category you describe that
as right now, there's so many different
things that are competing for that.
And so we want that space.
Yep.
Erica Rooney: Well, as
a user of Duolingo, I
Speaker 4: mean,
Erica Rooney: they
Speaker 4: get, what are you using it for?
I don't think it's taking.
Erica Rooney: It's not taking, I'll
tell you that it's not taking, but
they do get on you when you don't.
If you miss a day and then lemme tell
you if you said the post notifications 27
days, they really get on you, I guess said
Speaker 4: the streak
thing in the leaderboard.
I don't know how, how you're
doing from an adoption, but.
Yeah, it's really thoughtful.
Yeah, it's, it
Erica Rooney: is very,
it's done very well.
But Jesse, we love to bring
some light and levity.
See what I did there to the conversation
by having a fun little last chat game.
And this is where we pull out our phones,
uhoh, we get to share the context between,
what was that last chat we put in?
What did we use it for
and what did it give us?
Yeah.
We've had professional answers,
we've had personal answers.
Everything in between.
Jess Lipson: Yeah.
All right.
But it's great.
But my phone's off, but
I, I remember you did your
Erica Rooney: homework ahead
of time, not on the bus.
Like, you know, someone over here whatcha
Jess Lipson: talking about?
Erica Rooney: No, I, I don't name
names, but then we'll kick it off
with you, sir. You are prepared.
What was that last chat?
Jess Lipson: So I was driving over here,
uh, at the stoplight at Crabtree Mall.
A light popped on my truck and, uh, I took
a picture of it and I asked chat pt, like.
What is this?
What's the problem?
And I guess it was like some
seatbelt sensor in my passenger seat.
Bachi BT totally got got the answer for.
Hold on, I gotta ask
Erica Rooney: the question.
Did you drive the Maserati in here?
Jess Lipson: No, I drove.
I drive a truck.
He had a pickup truck.
Erica Rooney: Truck.
Oh, you said that I, somebody followed
me in with a Maserati today and
I'm trying to figure out who it is.
Yeah,
Jess Lipson: no, it's not me.
Okay.
I drive a pickup truck.
Um, just check it.
And then the one before that
was, uh, we're doing like a
president's club in Paris and.
Um, someone on our team was suggesting
like three or four hotels, and she was
like, one was like, this is in the first
district, this is in the 15th district.
And so I just like went to Shachi pt,
Hey, we're doing a president's club.
Would the first district
or 15th district be better?
We're looking at hotels.
And it gave me a great answer
and told me first district.
And so, uh, I was able to make
a better decision on that.
Um, so just wide, wide range of, uh.
Different things that I,
I've used it for recently.
Erica Rooney: I love it.
Greg, what was yours?
Nah,
Jess Lipson: I'm gonna go last.
Erica Rooney: Oh, mine's good.
You know, so I had the incredible
founder and Cee o of women tech
net court on my podcast, glass
ceiling links and sticky floors.
The other day, Anna RedSky.
And I said, I need like a fan girl post.
I said, 'cause I wanna blow this woman up.
She's incredible.
Uh, write me the most over
the top fan girl post.
And so they sure, sure as heck did.
And they did it because I've used
it so often for some of my content
production, they do it perfectly.
Same with the perfect amount
of emojis that I like, not too
over the top like it used to do.
No m dashes.
So, you know, that is the dead giveaway.
Um, and no weird call to actions
at the end that are just too quirky
because I've used it enough to do that.
So that was my fangirl chat GBT request.
That's the
Jess Lipson: amazing thing with chat GBT.
And I think why OpenAI, you know,
has a really good chance of running
away with the consumer market is
they've added memory and like.
It is just a great lock-in.
It's getting better and better.
You don't have to give it as much context.
If I can just say, Hey, like, is
there a good Mexican restaurant?
It knows I'm in Raleigh and, and,
um, just makes things so much easier.
And so I think that's
why, um, even though.
Gemini is probably just
as good of a model.
I think they've captured the
public's imagination and they have
so much momentum that I think it's
gonna be really hard to catch up.
Absolutely.
Erica Rooney: That's why I love these
posts that are like knowing everything
you know about me, chat, like, hype
me up and get me ready for the day.
Yeah.
And it will just bombard
you with a best answer.
Mm-hmm.
Speaker 4: Yeah, I mean, I think the.
I was gonna bring the same thing
up that, that Jess brought up, that
round memory, the switching costs
Speaker 5: Yeah.
Speaker 4: Are gonna be massive.
Like why would I leave it now
that it knows everything about me?
Exactly.
Jess Lipson: Yeah.
Right.
Speaker 4: It knows all my preferences
as much as we use Gemini in the business.
Speaker 5: Yeah.
Speaker 4: I've still
predominantly used chat GBT Yeah.
For things right now.
And I know the power of Gemini
'cause I was using its experimental
deep research last year.
The end of last year for a lot of things.
Mm-hmm.
Right.
And uh, but that is.
An amazing lock in.
Mm-hmm.
Right.
And so before I get to my chat,
like what, what is the thing, like
what do you, what do you believe
is defensible moat right now?
Mm-hmm.
For SaaS based software?
Yeah.
Because speed of competition or the
time to competition is nearing zero.
Jess Lipson: I think, uh, taste,
you know, design, taste, um, and
kind of product vision are, are,
you know, more important than ever.
So, um, and then I think that.
To the point we were just making
about chat, GPT, the more that you're
able to, um, train on specific data
that's relevant to the, the niche that
you're in, um, I think it potentially
allows the winners to get more of a
mode and more because the winners.
Um, get their models and get better
and better and, um, and then that
causes them to have more users, which
causes them to get better and better.
So yeah, I think it's, um, there's
definitely like getting out early,
being the market win the market
leader is probably just as important
as like back in those.com days when
we're talking about like eyeballs.
It's like if you can, um, make your
model and, and I'm not talking about.
Us or most companies creating their own
large language model, but it's more all
the training documents and prompts that
you can put around a model to make it
really relevant to a specific use case.
I think, um, that becomes the moat.
Speaker 4: Yeah.
Everyone that I'm hearing in, in
the VC community is saying that.
Kind of same thing, right?
There's a few different
things they talk about speed.
Mm-hmm.
Right?
Because it used to be
first, best or only, right?
Are you first in the market?
Are you the best?
Are you only, yeah.
Now I say now people like from a
large, from a foundational model
best is like relative in, in
terms of like every two weeks.
Mm-hmm.
Right?
People say, oh, open AI can't do math,
and they just won the math competition.
It took gold last weekend or whatever.
Yeah.
Right.
But I think the, uh.
They, they talk a lot about speed and data
and to your point, people that can get
very deep in a specific vertical, right.
Or industry.
And then you lock in the data
because I'm real bullish on
service as a software right now.
Absolutely.
And I say the, the four pieces of a
defensible moat in the service of a
software world is your reputation.
So personality led growth,
your relationships, so
community led growth, right?
Your clients and partners.
The data that sits underneath.
And then the services you wrap around it.
Mm-hmm.
Right.
And so from a SaaS perspective,
you can get them locked into data
or you can go deep with that fine
tuning as as you're describing.
Yeah.
Right.
And have all that context.
Like, so imagine being like the,
like Harvey is doing for legal.
Yep.
Right.
They've gone so deep there.
Um, I do think that the folks that
are coming in late to the party
are, it's gonna be challenging.
I agree.
I say we're in an era of taste and
trust is what I always tell folks.
Yeah.
Right.
Because there's gonna be so many
options you're gonna go with, you
know, like c Claude, like, who gets me.
Mm-hmm.
It's about vibes.
Right.
Vibes and vision.
Mm-hmm.
Is what Jess is saying, uh, to us.
But, all right.
So I've, I've, uh, delayed enough here.
Let me see which one I wanna, uh.
I have one, but I'm like
deep in prompt engineer.
I'm a point, I have 25 certifications,
so I like really nerd out on this stuff.
So I'm not gonna read a
whole page of, of a prompt.
So I'll read a shorter of C. I'll
read a shorter one, which is a lot of
times people will gimme a PDF and it's
like, okay, like a deck or something.
I'm like, all right.
Let me do this better using gens,
spark, or using gamma or something else.
And so someone sent me A PDF and I
didn't really like it, so I said, all
right, read through the attached PDF,
extract all the content from each
slide and put it into a gens spark
prompt for me to rebuild this deck.
Slide by slide.
Nice, right?
And so I took their deck, ah, made
it better and I made it better than
Jess Lipson: one that you actually wanna
Speaker 4: read.
Correct.
And one that I wanna pass
along to other folks.
Oh, okay.
Nice.
Right.
Because I was like, I can't give this
to the audience I want to give it to.
Right.
Their expectations are
higher, more thoughtful.
Right.
Also because I'm using these
tools, to your point, it has the
memory on me and knows my voice.
Mm-hmm.
Right?
So it puts it in the tone
in which I would present.
Yeah.
And it gives me speaking notes
and all these other things.
Right.
Yeah.
So I, I use it.
I use it a lot of different ways
from writing code to vibe coding,
to making music, you name it.
So, yeah.
That's awesome.
Erica Rooney: I wanna wrap
it up with just one question.
Yep.
If you had to sum up your AI philosophy
in one sentence, what would it be?
Jess Lipson: I would say, um,
it's important to.
Use the tools a lot.
So I, and this is more than the
sentence, but, um, I've actually been
getting back the last nine months
into coding and I was just coding at
breakfast this morning with Claude Code.
Um, and so I think the,
the advice is like, um,
you really need to constantly be
using the tools and evaluating them,
um, because, uh, I think the biggest
mistake you can make in AI is.
Oh, I tried to implement
that and it just didn't work.
And then, you know, waiting a year
before even thinking about it again,
it's like, I feel like pre ai, that
approach would work of like, oh, we
tried to implement this type of software.
It failed so.
Maybe we'll come back
to that in five years.
Whereas I think with AI it changes
so fast that you need to always be
reevaluating your assumptions and
like something that didn't work a
month ago might work really well.
Uh, today.
It's like AI was horrible at dealing
with images three months ago.
Now it's gotten a lot better and I bet
it's gonna get a lot better next month.
And so you, um, I think I
would say just like always be
paying attention to the news.
Uh.
It's like a year happens in two weeks.
Um, in terms of the, the pace of
change and, um, and so don't, I
think the always err on the side
of trying to use AI too much versus
airing on the side of dismissing it.
Which I think is like the most
dangerous thing you can do.
So
Erica Rooney: try and try again.
Yeah.
I love it.
Jesse, thank you for
joining us on the podcast.
Absolutely.
My pleasure.
Oh my gosh.
Thank you.
So fun.
Yeah,
Jess Lipson: this was great.
Erica Rooney: Thanks for joining
us on AI, voice or Victim.
If you want to stay competitive
in the AI age, start now.
Take one insight from today's episode and
put it into practice in the next 24 hours.
Make sure to follow us.
Share your thoughts and subscribe
for more actionable AI insights.
See you next time.