AI Meets SaaS with Jack Altman, CEO of Lattice
Learn about AI's impact on HR in the SaaS industry. Hear from Jack Altman, CEO of Lattice, about balancing technology and people and his view on the most exciting AI innovations in the market.
Hi, everyone. I'm Nick Meida, CEO of GainSight.
I'm excited to welcome you to the next edition of our new web series, AI Meet
SAS,
where we talk about how this incredible world of AI innovation
applies to softwares and service companies,
where there's unique questions to answer around pricing, roadmap,
customer buy-in around privacy and security.
Each time I'm featuring somebody I really respect,
let's think about AI in their own business.
Today, I've got Jack Altman.
Jack is the CEO and co-founder of Lattice.
By the way, we use Lattice internally.
He's a huge fan.
Lattice is leading People Success platform,
which I love that term since we're customer success,
that's on a mission to build cultures where employees and their companies
thrive.
So super cool mission.
Great to see you, Jack.
Thanks for having me.
And we also love using GainSight.
There you go.
I got it right back.
Mutual of fast, I love it.
So we always like to start with a nice breaker.
So I'll say that I think Lattice just turned eight this month, which is pretty
cool.
Do you remember anything about you turning eight on the personal side years ago
So I'm pretty sure this was an eighth birthday because I did a few of these.
But I loved bowling when I was that age and I did a bunch of bowling birthday
parties.
And I actually, I think when I was like 11 or 12,
I was even in a little bowling league, which is like a ridiculous thing to
think back on.
But I loved bowling.
But I thought a bowling birthday party when you're like eight years old,
it's like an arcade and cake and there's bumpers because you're eight and it's
awesome.
I love it.
I was going to ask about the bumpers.
And I actually still like bowling.
It's fine.
I'm not good at it, but it's fun.
It's a good place to hang out.
Cyber is great.
Yeah, no shame in bumpers.
I love that.
Some of my adult friends still use bumpers.
So OK, so we want to talk about AI.
I'm actually going to turn the first question over to something who's
our head of product management for AI GainSight.
Thanks, Nick.
Hi, Jack.
So my first question for you, Jack, is how has AI
impacted tech in the HR industry?
And more specifically also, how is Lattice using AI today?
Yeah, well, something that we've done for a while on the analytics side is
using it
for sentiment analysis.
So Lattice collects a ton of data.
And a lot of that is in written form.
And so on, things like an engagement survey or something like that,
you can get specific answers in a dropdown,
but there's a lot of free form.
And sentiment analysis is like a really great tool that is quite accurate in
most cases.
Not perfect, but it's quite accurate to basically just give you general pulses
across a population.
You wouldn't want to use it for a particular piece of text
necessarily, but across populations, it can be really useful.
I think in general, one of the first places
we're going to start to see this current wave of AI really matter
will be probably on extracting insights from the data that
exists in HR systems.
And so this is one example, but also crossing different sets of data
around your people and pulling out interesting insights or flagging data
that was otherwise hard to get.
I think AI is really well suited for that.
Another area where I think AI is well suited for,
but I don't know that we know yet how it'll get implemented is text generation.
And so when I chart, for example, one of the obvious concepts that comes up,
which to me is a very open question, is, should we have AI generated
performance reviews?
To be totally honest about it, it kind of gives me the heebie-jeebies
because my own experience as a manager is that the process of writing itself
is where so much of the value comes in the performance conversation.
The writing of it is a piece, but it's the thinking and the distillation
of what's the advice, what's the feedback, what's the recognition I want to
give somebody,
and then going and having that conversation.
And so that doesn't mean that I'm particularly against it.
It just means that I think as you consider concepts like this
and this applies all over the place, I'm sure there's applications or analogs
in a gain site context, you just have to be careful that you aren't using AI
to remove the human parts of the loop that are extremely important
and instead are using it to support and further distill the parts that are most
important.
So in performance reviews, if there's a way to use AI to spend more of your
time
thinking about what matters most and having a better conversation, all for it,
but just replacing the human part of the conversation,
I don't know that that gets us where we want to go.
I love that. I've thought a lot about that because obviously a lot of users,
and I do my reviews.
And I always feel like the reviews, there are a lot of works,
it's sort of you almost procrastinate them to the last minute.
And then I do them and they're so incredibly valuable.
But I'm with you, Jack, that there's a human element where it forces me to
think about
when I'm really writing and think about my team and what's most constructive
for them.
So I do think there's some danger in just totally outsourcing it to the AI.
So the other thing I found, by the way, just as a lot of users is like,
we have 1,300 people in the company, so I don't know how many comments we get
in our service,
but let's say it's 5,000.
And so I'm assuming this is what you're seeing, too, Jack,
which is very high volume data.
I'm not going to click through, maybe I would every single survey,
but I use the sentiment analysis and things like that to just understand what
the big trends are.
So I'm guessing that's the type of use case you're seeing, right?
Yeah, definitely. I think that's exactly it.
And yeah, it's sort of like, can the AI go to the gym for me?
It's like, well, they could, but you got to go to the gym.
I think one other area where AI will matter in HR tech,
but I actually think this is going to just be an all software,
will be helping users get data in and out of systems more easily,
and helping configure the systems more easily.
So you take a product like Gainsight, which is incredibly powerful
and incredibly rich.
It has a lot that's involved in setting up the tool in the way that a user
wants.
And so a product like yours, and also, I think a product like Lattice's,
which is also a heavily involved product,
you could imagine a world where an HR leader or a customer success or BizOps
leader for you
was able to use natural language as an interface to say,
hey, I want the performance for you to be set up this way.
I want, you know, and a new employee onboarding for our HRIS to look like XYZ,
and I want these things to happen in these cases.
And then it goes and configures that system.
I think this will be one of the things also that I think AI will be really well
equipped to do,
is make these systems much more usable.
It's so interesting.
It's funny because science and other smiling,
because that's one of the big things we're working on,
which is like a setup assistant for that exact exact example.
We demoed some of our conference.
So we definitely are thinking alike.
This is a little bit off-script, but I'm curious, like the chat UI,
chat as a UI, you know, does that, you know, obviously we're getting more used
to that with,
you know, consumer applications.
Does that become more of the UI for SAS and like, what's the mix of traditional
you know, menus and dropdowns and filters versus chat?
I think we'll, we're going to all learn together where the boundaries are.
I think a point and click has been a very,
you know, it's been a very good UI as well, like GUI's work.
So there's a lot to be said for that.
I think natural language is going to work in more than zero places.
And right now you can basically use it in the way we're talking about in zero
places.
So I don't think you'll use it in 98% of places,
and I don't think you'll use it in zero percent of places.
So I don't know what the boundary will be,
and I expect that it will have some, you know, S curve adoption at some,
you know, and get up to some rate that I don't know what it is.
I sort of suspect, if I had to place a bet now, that it will be more important
on the admin side of these kinds of systems than on the user side.
And that really what you'll want to do is get away from the heaviest usage
situations, but on the light usage situations,
you just want a good product to make the, you know, the input output formats
we're all familiar with work.
But we'll see. But that would be my guess.
It's a great thoughtful answer. Yeah, I like that a lot.
So you alluded to this idea in performance reviews of
there's some things that can be automated, but some parts of it are really
understanding
that human on the other side, and really understanding like their goals,
and also the value of the company and where they're trying to get to.
So you probably know, is like a game site, one of our things we talk about is
human first,
our company purpose is to be living proof.
You can win in business while being human first, very, very deep in our mission
deep in the water, a game site.
And I read something recently where you, you know, you talked about some,
how people are thinking about concerns around AI replacing people,
which if we're being real, is a concern a lot of employees have.
You know, what do you think? You know, where's that can happen?
Is there an S curve of replacement that's going to happen?
How do you think about that?
Some jobs that are done today will get replaced.
That doesn't mean that all the people who are doing those jobs have to get
replaced.
And this is also why like societal change or technological change happening
very rapidly
just leads to a lot of like structural displacement that just like takes a lot
of
years to work through because people need to get retrained and things like that
But you think about the jobs happening at game site today, like how many of
those jobs existed
10 years ago, 30 years ago, 50 years ago, like zero, right?
I mean, at some point you get back to the number zero.
I guess sales is kind of, there's some things that are.
Yeah, some things were there, but most weren't.
Yeah. And so like, you know, that we're going to go through versions of that.
AI, you know, with or without AI, that is happening.
It's just that what's challenging with AI is that the anticipated speed at
which that's going
to happen is going to be so much faster.
And so there's a there's a very valid fear that, you know, overnight something
could happen that,
you know, drastically reduces the need for humans in a particular area.
What I would say though is you it's really that while it will likely happen in
some areas,
A, it's not it shouldn't be the case that total employment goes down in the
medium term,
even if it doesn't the very short term.
But hopefully there are more jobs created as a result of this, like there have
been from the
internet, which also replaced a lot of jobs that no longer exist.
But the other thing is that it's really hard to predict what's going to do what
An easy example of this.
So like, you know, co-pilot software engineering, people are way more
productive.
Like once people figure out how to use co-pilot, they're just more productive.
Totally.
So what happens in a world where all engineers are twice as productive?
Do you think we get half as many engineers?
No, definitely not.
We're going to have way more engineers because it's going to be more valuable
to companies.
There's going to be more demand for this more productive asset.
Or at least that's what I think.
And I think that has been like born out over time as engineering has gotten
more
productive.
We haven't gotten fewer engineers each year.
We get way more.
So, you know, it's hard to call it.
There will be some amount of this.
And I think we all need to be thoughtful as we go and through it together.
But like, I think just like having a general, you know, I'm a general techno
optimist,
whatever the right term is for that.
And I sort of believe that valuable steps forward are going to be good for
everyone.
I love that.
It's a really good point.
And there's some nuance about the medium term and the short term and things
like that.
So really interesting.
You know, one of the other things you wrote about actually, you spoke about at
SaaSter,
it was really cool post if you haven't seen it, which is some misconceptions
about SaaS.
And one of them, which is interesting, you know,
probably counterintuitive to some people, I'll just read it,
says, "Misconception number five, we should invest to avoid tech debt that will
cripple us later."
So how do you think about tech debt?
And then how does it connect at all with AI?
Well, what I meant by talking about, what I meant by we should avoid having
tech debt is
it's not avoiding debt isn't free.
Avoiding any kind of debt doesn't come for free.
And when you're an early stage company, like it's a good time to take on debt
in order to
increase the speed and likelihood of getting product market fit, which is, you
know,
the binary yes/no, you have a company that exists or you don't.
And so given there's that binary moment of either we're going home or we, you
know,
live to see, you know, another day, you want to do basically everything you can
And so I think it's not that you want to go out of your way to incur tech debt,
but
it's worth it to take some shortcuts at the beginning because odds are the
things that
are going to work.
There's a good chance that what you're building is you're going to pivot away
from anyway.
And so, you know, you might want to give yourself a chance to have four shots
on goal for different
product areas and you want to build them as quickly as possible.
And those aren't always in the cleanest, most long-term durable way.
And the truth is you know, it's painful later, but it's a total champagne
problem to when you
have hundreds of engineers go back and, you know, spend a year paying down tech
debt with
a portion of them.
And so that was like the nature of that comment was just that it's not
something to be
scared of. It's actually a tool. It is a tool to use like financial debt that
can be a big asset.
And then how do you think about AI? Because obviously we're all thinking about
how to innovate
and AI. Most of us don't really have a clue what customers are going to want in
a long term.
Should we embrace the same mentality of not worrying so much about tech debt
and focusing on trying things?
I think we should. I think it's easier for smaller companies to embrace trying
things than big
companies. I know you've read Innovator Dilemma too and like, you know, you
probably are feeling,
you know, in the middle of a small startup and a big enterprise kind of like I
am. And you know,
you see both sides of it where like, "Can we really divert resources to this
thing that's not
going to work?" But also like, "We've got to keep trying and there's so much
edge here if this does
work." So, you know, aside from the should we, there's just like who will. I
think in a lot of cases,
it's really important that companies like GainSight do innovate. Because it
seems to me that in
more categories than not, it's the GainSight stage companies who are most
likely to benefit most.
Because you already have the distribution, you already have the workflows, you
already have,
you know, the customers who want the thing. And so now you can add AI into
something like
GainSight. That's a much easier hill to climb than starting from scratch and
building GainSight
and doing AI. And so I think if the midsize companies can muster the ability to
do AI,
this is a type of trend that will benefit those companies greatly. So I guess
to the
should point, yeah, I think most should. And I think particularly most mid
stage, you know,
not that obviously I know GainSight is late stage, but midsize compared to like
a Microsoft or something
like that, should be working hard to do that. Because that's, if you think
about it logically,
like that's where so much the value ought to accrue. Yeah, it makes total sense
. Yeah, it's hard
for the young companies because they don't have the data and the workflow.
Although I'm sure some
of them will thrive. And it's harder for the big companies because of innov
ators to lemma. So
I, by the way, compared to Microsoft, I think we're still a tiny, tiny company.
So it's all like an
eight year old. It's like a bowling. Exactly. So okay, but speaking of that, I
'm going to turn
it back to Shonton around like innovating in AI. Right. And Jack, so who have
you seen that's been
doing some of the most innovative work in AI in recent times? Well, there are,
I'll give an example
that's like maybe talked about less. I think obviously there's a lot of really
cool like,
there's a lot of cool like companies building models that we all know. There's
a lot of like
generative AI companies that are like specific applications on top of those
models building things.
But something I have found really cool recently is a company that where I know
the founders well
who are relatively early stage like, you know, like series A stage and their
product itself has
nothing really to do with AI. But they are using AI to the fullest extent to
run the company that
I've seen from any other company that I can think of where they're using it to
not just like have
internal tooling, but they're using it to do sales and they're using it to do
support and they're,
you know, finding all of these ways to make use of AI products to build the
most efficient business
possible. But what's interesting about it is it's not necessarily just small
edges.
There are certain business models that will become possible that weren't
possible before
because they didn't work when you needed to scale them with humans. But if you
can scale them with
AI, the math can work all of a sudden. So, you know, you brought up SaaS or
something like Jason
Lumpkin has talked about is like, you know, depending on your ECV, you just
like can't fund
sales teams. Like it just doesn't work to fund a sales team against a $2,000 EC
V.
But if you have a very inexpensive AI bot that is able to do quite a lot of the
sales,
even at 50% the effectiveness, but at 2% of the cost, maybe a 2K ECV company is
possible. And so,
maybe new products will exist not because the AI let us build the products, but
because the AI
let us distribute the products. And so, I think we're going to see some
surprising things where
it's just like hard to call it right now because it is applying to so many
cracks and
crowdsets of businesses right now. And it's like, you know, there's just a lot
that could go from
here. So, that's been a really cool thing for me to see. I love that. And it's
kind of the next
way where the internet allowed distribution that wasn't possible before and
there's certain x phase
there. So, it's sort of a good setup to one of our last questions, which is,
how do you think
about that post sales experience at Lattice? You guys really focused on that.
It's also,
I'm sure, you know, a lot of people and a lot of, you know, people working
together,
workflow and all that. Like, what are the big opportunities in the post sales
world run AI?
I think there are at least a cut. There are probably many, but at least a
couple that I can
think of. One is products like Gainsite help a huge amount. And even with that,
it is still
so messy to find all of the data about the customer that is relevant. You don't
know who knows who,
you don't know what exec has talked to, which exec at the customer, you don't
know what exactly
all happened in the presale. You don't know if the champion was a user of ours
from a different
company that they worked at. There's just like all this data and a lot of it's
useful. And
finding all of that data and surfacing what I should care about, I think is
something that is
on top of a platform like yours, I think could really elevate what's possible
for
a customer success person or just for a company. So I think there's going to be
something around
like getting the real totality of the picture, you know, like those moments
where you're like
trying to do a renewal and you realize your head of people is best friends with
their head of
people and you're like, oh, I wish we knew that three months ago. You know, it
was just like
stuff like that that happens all the time. I do also to this other point, I
think a lot of
a lot of customer success looks a bit like what I was just describing with, you
know,
that last example where it's not that you're, you know, going out and selling
them hard or you
can replace the human touch with a bot, but there is a process to these things
to be followed.
And when you've got a small team with a lot of accounts, trying to track all of
these in systems,
balls just get dropped. There's just like no way around it. Like, you know,
like Lattice has 5,000
or 5,500 customers. There's just no way that all of those, there's no way that
if we fully
inspected, you know, all of those that we would be happy with every single one,
just, you know,
it's the same thing on the sales side on the pre-sale side too. There's no way
every lead is
being maximized. But with AI, it's possible that you can drastically close the
gap of
what we aspire to be doing with each customer and what really is happening
through things like
chatbots, maybe in many cases nudges to the rep to like get them on calls at
the right time to
maybe suggest calls that, you know, we wouldn't have thought of things like
that. So I think
there's a ton there. I love that. Yeah, it's funny because we have a, since you
're a customer a few
weeks, we have a feature coming up called cheat sheet, cheat sheet, which is I
think the beginning
of what you just talked about, it kind of builds an exact prep based on all the
dating gain side.
It's actually pretty incredible how accurate it is. You know, I call it 95%
accurate.
Shump it better than it's better than not having it. Exactly. So it's a shot
that I went
few weeks from now. I think it's shipping, right? Yes, November, mid of
November is when you'd have
it. There you go. Try it out. So want to get everyone out on time, including
Jack. So thank you so
much. It's been incredible. I was like to end with some word association just
for fun.
So three things. The first one, I think, is more tidy personally. So super
smash brothers.
Tell me what that brings up to you. Yeah. I don't, you know, I think I had that
as like my
a fun thing about me. And it just sort of like stuck somewhere. And so I do
love that game. I
played that game so much as a kid. And I did get, I got good at it. I really
got good at that game.
And I loved it. It's been too long. But I do want now I've got two little kids
now.
And so between that and work, I have not had the video game hours that I want.
But I do dream of the future where I'm right back in it. I love it. Human first
. What's that mean to you?
Oh, I think I mean, it, it's sort of a, it's a way of to me, what comes to mind
is it's a way
of building companies. And it's sort of like one of my driving principles when
I think about building
a company, which is I believe that people are the most important input to a
successful business. And
I think a, a people first company is going to be more likely to have a winning
business. And so
that's just sort of, you know, when I think about are you, you know, you're,
you got your people,
you've got your, your customers, you got shareholders, I believe the center of
that flywheel is people.
And then everything else falls from that. Love it. AI. Last one.
I mean, you know, it's funny in the context of human first, but I'll, there's a
lot you could
say about it, but just to put it next to human first. My hope is that AI is
used in a way that
brings more humanity forward and that it takes away the things that people don
't want to be doing.
And it lets people spend more of their time when the things people do want to
be doing.
You know, one of the things I was thinking about, as you were asking about tech
debt,
is it would be an emit, like engineers like working on a lot of things.
Sometimes they like
working on tech debt, but probably less often do they like like cleaning up
tech debt from six
years ago than building a new product or, you know, something else that's more
interesting.
And so I think my hope is that there's a lot of ways that AI does a lot of the
stuff that is,
you know, less desirable for people to spend their time on. And so I hope it
leads to even more,
you know, human first companies. Love it. Jack, this has been amazing.
Learned so much myself. I'm sure the audience does did as well. And I look for
the future,
whereas your kids grow, you can go bowling with the bumpers and play super
smash bros with them.
So till then, we'll figure out AI and thank you so much, Jack. Thanks,
Shantan. And thanks to everyone for watching. Sounds great. Thanks for having
me.
Okay, great. I think we had a good that was awesome, man. Thank you so much.
My pleasure. Funny. Like, obviously, we didn't talk about this live before,
but gosh, the alignment's amazing. That really was great. Super natural. So
definitely.
We need AI alignment. That's key. So, yeah, I'll make it. Yeah, totally agree.
The comment on game size companies being the new AI startup subsets is
unpointed. Throw that on the website. I actually think it's true. The caveat is
that I think
the big companies, the real big companies ought to have the advantage. I just
don't believe that
they'll be able to, or it'll be much harder for most of those companies to turn
the boat.
Although the Fang or whatever they're called now, they're doing a surprisingly
good job,
but the next rung, we'll see. But yeah, the game sites should clean up. It
should be very
hard as a startup to do AI game site. Love it. All right, Shantan. Keep going.
Exactly.
Awesome, man. Okay. Anything I can do to the replay of the favor, let me know
and we'll talk to you soon.
Sounds good. Thanks so much. Bye. Thanks, everybody.
[buzzing]