AI Bubble or Breakout: The Reality of AI in Sales

AI Bubble or Breakout: Are you building real, needle-moving workflows with AI, or collecting “bubble” use cases that create overlap and extra change management?

In this episode of AI Powered Seller, Jake Dunlap explains why the “AI bubble” is not AI itself, but the noise on top of it, including duplicate tools, overlapping features, and organizations investing without clear ownership. He breaks down how GTM leaders and frontline sellers can evaluate AI the right way: start with the revenue bottleneck, then choose technology that solves it.
Innovative Seller

You’ll also get Jake’s tactical seller playbook for adopting GenAI in a sustainable way:

• Identify your three most painful tasks
• Refine AI outputs so they sound like you
• Build a simple ongoing roadmap so you don’t try to solve everything at once

Check Out: 

• JourneyAI (free trial): https://meetjourney.ai/
• Jake Dunlap – https://www.linkedin.com/in/jakedunlap/
• Skaled - https://skaled.com/

If you got value from this episode, share it with a teammate in Sales, RevOps, Enablement, or Marketing, and make sure you’re subscribed so you don’t miss the next drop.

Episode Transcript

SPEAKER_00 (00:00):
If we fast forward a year from now, are you gonna be

(00:02):
cleaning up a bunch ofbubble-based AI use cases that
never move the needle?
Or is your organization gonna beheading into 2027 looking
completely different and workingcompletely different?
Are we solving real businesseschallenges here that are truly
the bottleneck within our org orwith our customers?
Yes or no?
How do you make sure that you'reimplementing in your day-to-day

(00:23):
breakout use cases and notbubble use cases?
What's up, everybody?
Welcome to another episode ofthe AI Powered Seller.
I am your host, Jake Dunlap.
Today's episode AI bubble orbreakout.
So today we're gonna talk aboutthe hype.
We're gonna talk about what'sreal, what's not.
You see companies likeMicrosoft, they are doubling,

(00:44):
tripling down on theirinfrastructure.
Amazon is going all in with thenew custom silicone, which I
don't even know what that means.
Google just released the Gemini,I think it's called the 3.0
model, which is insane, by theway, if you haven't tried it.
Nvidia keeps crushingexpectations.
OpenAI is just continuing toroll out enterprise features.

(01:05):
So you have all of theseorganizations, these companies
that are valued at trillions ofdollars in market cap continuing
to double down, continuing toshow tremendous potential.
But what does that mean for allof you as sellers?
So today's episode, I'm going tobreak down is this more bubble
behavior, or is this morebreakout sustainable behavior?

(01:29):
Because I think we all agreethis isn't going anywhere,
right?
AI is going to be here.
It's going to be something thatwe're using for in some some
form or fashion for quite sometime to come.
The key is all these othertools, right?
All these different AIgo-to-market tools.
You know, I can't even tell youhow many new tools pop up where

(01:49):
a go-to-market leader's like,Jake, have you heard of this
one?
Or have you heard of this one?
And I think the the issue we seeis because we see all these
breakout AI use cases, we thensay, well, oh, these RevTech
companies, you know, could theybe the next X?
And for most of them, the answeris no, right?
But the reality is, I think wehave to try as go-to-market

(02:10):
sellers, as leaders, to stay atthe forefront, but also try to
break down what's actually goingto drive impact because that's
the key to AI.
The key to AI is not thinkingabout, hey, here's a cool new
tool that can do X.
Should we do X?
It's what's the issue I'm havingin my business and can this tool
solve it?
This episode is going to breakdown the real AI bubble.

(02:31):
What's inflated?
What do I think is real?
Because there is a lot of verylike real potential here.
And why go-to-market teams andfrontline sellers have to start
to invest time to learn thesetools, become experts, and make
sure that they can stay at theforefront of what we know is
going to happen and spend lesstime worrying about, you know,
maybe some fringe use case.

(02:53):
So, with all that being said,let's jump into it.
AI powered seller.
All right.
So, first let's talk about whatthe AI bubble actually is.
Right?
What is what is an AI bubble,right?
Is it overhyped?
Is it doomed?
To me, when it comes to AI, it'snot that there's a bubble,
right?
For for those of you, I wasn'tin the workforce when you know

(03:14):
the internet bubble happened,for example, where all these
companies raised a ton of money,you know, it's all the dot-com
boom, and then there was thishuge, you know, uh reckoning
where all these companies wentout of business.
But what didn't change is thatthe internet changed the world.
And I think that is where we areat now when it comes to AI, that

(03:37):
it's not that AI isn't the nextthing.
It's all the noise sitting ontop of the real capabilities.
It's all the noise of like thisthing's AI, this thing's AI,
this thing's AI, to where, youknow, potentially we might see
and probably will see some ofthese companies go the way of

(03:57):
pets.com.
Shout out to pets.com.
If you don't know that story,it's a good one about you know,
a really overvalued company thateveryone thought was gonna be
huge and then tanked.
So for many of you listening,understand this AI as a
technology, generative AIspecifically, is not going
anywhere.
So investing in the tools,knowing how to use them, finding

(04:17):
new ways to use them in yourday-to-day is gonna be required.
And so when I think about thebubble, right, this is a really
good example.
I talked to a CEO and he toldme, Jake, look, we invested in
three AI tools and none of themworked.
Okay.
And when I asked him, I said,okay, well, great.
Whenever you were evaluating,when you were thinking about
these technologies, what was theworkflow or bottlenecks in the

(04:40):
revenue engine that you weretrying to solve?
The answer, yeah, that's a goodpoint.
And I think with a lot of AI usecases, that's exactly what we're
we're doing.
We're saying, hey, look,intuitively, this tool sounds
kind of cool.
But guess what?
If we're implementing a toolthat solves priority number 32
and is a big change managementlift, it's probably not worth

(05:01):
it.
And so for a lot of you outthere, when we think about how I
can make sure I'm investing mytime and energy in breakout use
cases, it really all goes backto the issues or areas of
opportunity within the revenueengine, from how we generate
leads, work with customers, howwe actually go and grow our

(05:21):
current customer relationships,making people more productive,
et cetera.
And so where I see the bubble istools that have identical
features.
I cannot tell you right now, andI'm sure many of you see this,
how many RevOps teams I workwith that have invested in one
tool maybe a year or two ago.
That tool now does five things,but now they've layered in three

(05:42):
other tools that did, you know,this feature set when this
company didn't have that featureset.
And so for a lot of my AItechnology investment
individuals out there, what Iwant you to think about is are
you staying up to speed with howthe tools you already have are
using AI, or are you layering intools that have 20, 30, 40, 50%

(06:05):
overlap just because the toolsare evolving?
And I think for a lot of GTMRevOps and enablement leaders,
it really is critical tocontinue to go back and
understand what the tool stackis that you have today.
And even for my reps out there,if I was a rep, I would
constantly be paying attentionto the roadmap of my sales
engagement platform or my callmonitoring platform.

(06:26):
So this isn't just about RevOpsor an enablement.
If you're a rep, you need to bepaying attention.
So if your team's not payingattention, you can still go and
capitalize on those features.
The other big issue I see is noownership, is that there's
nobody really internally owningthe AI roadmap to see the
redundancies or the issues withthe workflow.

(06:46):
Again, we do a lot of this typeof work for companies because it
is very difficult to dointernally, where marketing is
investing in a technology ortools or data set, finance is
investing, sales is investing,customer, everyone's investing.
And again, not only is thereredundancy, there's just a ton
of overlap in people doing verysimilar work.
Whereas if you have, and I'm allfor, by the way, a decentralized

(07:10):
slash department by departmentrollout of Gen AI.
That's what we preach to ourclients.
But you have to have acentralized group that's paying
attention to all thetechnologies being used, all the
new agents and assistants beingused to speed up development
time because you're gonna see alot of different groups creating
flavors of the same thing.

(07:31):
So if you don't have that kindof centralized workflow, you're
gonna really, really struggle tomove as fast as you can and not
have just a ton of overlap.
And so to me, those are the twobig ones that as organizations
are trying to say, Jake, look,we know we need to invest in AI.
We want to make sure we're notinvesting in the bubble tech,
but the breakout tech.

(07:51):
That is really what I try tolook at is I say, okay, are we
solving real business'schallenges here that are truly
the bottleneck within our org orwith our customers?
Yes or no?
Do we have a centralized placewhere you know we can look at
best practices of what peopleare doing, allow departments to
yes, they know their businessbetter than you know, IT, for
example.
So we do those two things, andthat helps to make sure

(08:14):
structurally as an organization,we are continuing to implement
things that are going toactually move the needle for the
business.
So that's the big first piecehere.
And what I want to do with therest of the same time, getting
very, very tactical into whatfrontline sellers need to know
and do about the bubble.

(08:36):
And as usual, everyone, if youare enjoying the episode, please
make sure to like, subscribe ifyou're watching us live on
YouTube, uh, if you're listeningon your favorite podcast
platform, make sure to sign upto get the alerts and downloads
when those come out.
Definitely share some of thesetidbits with your team.
That's what I always encouragepeople.
They'll say, Jake, what do Ineed to do to stay on top of
things?
I said, Look, you know, myself,the team, we're doing a lot to

(08:57):
make sure we're staying at theforefront of what's happening in
AI.
So, you know, share this littleclip with your team, maybe your
RevOps team, your sales leader.
I think any of them are gonnaget a ton of value out of it.
So without further ado, let'sget a little tactical.
Let's talk about what we need tobe doing in the trenches around
this.
I gave the example a littleearlier of staying on top of
your own product roadmap and theproduct roadmap of the tech

(09:20):
stack that your team is investedin.
I think that to me is probablynumber one that if you are a
seller, your team is investing alot of money.
I can't remember the last numberI saw.
I want to say it was maybe$800per rep per month in tech,
something like that.
If you include CRM and all theother things happening, so

(09:41):
you're talking about you know,companies investing tens of
thousands of dollars in everyrep around the tech.
Now, again, the issue is manytimes your own internal team
can't keep up with the updates.
And so that's what I wouldencourage all of you to do is
step one, become an expert inthe tools that your organization
already has.
I do believe many of you, asreps, should be getting

(10:04):
certified in those tech.
You know, you can imagine repnumber one really knows the
nooks and cranny of thisAI-powered platform.
Rep number two is hey, I wastrained to use it this way.
This is what I was told to do.
And they're just going throughthe motions.
I have to imagine that repnumber one, even if they're not
quite as good, is gonna get aton more value out of the

(10:24):
platform and hopefully moreresults based on what it was
supposed to drive.
And I think nothing illustratesthat.
I was talking to a rep who weretalking about AI, and what he
mentioned is that I feel like AIslows me down or it just it's
just not good enough.
So I need to do things manually,right?
He was using it to, you know,write me a cold email, tweak

(10:45):
this email before I send it out,etc.
And what happened is literallywith a couple of tweaks around
how he thought about prompting,how he thought about what the
role of AI is, which isn't to dothe job for me, but to get me to
V1, immediately came back andsaid, I was a complete skeptic.

(11:05):
I'd been using it, notimpressed.
He's I'm all in now.
What it took to get him to allin was him realizing that the
tool wasn't the problem.
It's the workflow and the waythat he was using the tool that
became the problem.
And I think for a lot of peoplewho are AI still step one, I
think that is probably outsideof staying on top of your own

(11:26):
tech, that behavior ofunderstanding that the tool can
do a lot for me, but its jobisn't to just do it for me, it's
to be a collaboration partner.
And so many reps don't need moreAI tools.
What they need is a workflowthat removes the real effort so
then they can turn their brainon and then get this thing to V3

(11:49):
or V4 in the same amount oftime.
And so if you're a rep outthere, you're expecting AI to do
your job for you, or you're arep who says, Well, actually,
this AI output's pretty good.
I'm just gonna copy and paste itand send and ship it.
Both of those use cases are notsustainable.
I just want to be straight up,right?
Think about those use cases.
If you give up on AI or youjust, you know, you know, say,

(12:12):
hey, I've got this thing, it'sworking for me.
There's going to be people thatare gonna outpace you.
They can just be moreproductive.
Like it just is what it isbecause they know how to use the
tools.
On the flip side, if you'recopying and pasting AI, well,
guess what?
Pretty soon the AI is just gonnasay, Oh, I'm already plugged
into this workflow to sendemails or whatever.
I'll just generate it and sendit.

(12:33):
And so when you think about atthe rep level, I want all of you
to think about your gen AIworkflow.
And and I'm gonna give you kindof my step one, two, three for
how to break down where youshould be using AI and how to do
it.
And we'll try to go into acouple of very specific use
cases for each one.
So the first is to identify yourthree most painful tasks.

(12:56):
So I'm talking about those kindof tasks that you dread, right?
The ones where you wake up inthe morning, you look at your
calendar, you go, Oh gosh, Ihave to do that.
Or over the weekend, you'rethinking through it and you're
like, oh my gosh, this is gonnatake forever.
So identify your top three.
And those could be prep related,they could be creativity
related, it could be, hey, I'mnot really sure how to handle

(13:16):
this, and maybe that's where AIcan help me.
It could be efficiency, meaningyou know, there's a lot of
different ways.
So some of the automations andthings that we're building for
clients are can it just look atmy calendar invite, look at the
domain, and automatically run anassistant that does my account
prep?
Yes, is the answer.
So it's really just looking forthose different ways that you
can alleviate your biggestpains.

(13:38):
So, step one for any rep outthere, what are your three
biggest pains?
For most of our clients, I willtell you the first area that we
eliminate for reps when it comesto like let's use AI to
eliminate the pain is accountprep, prospect-based research.
So coming up with snippets ofsub-industry trends.
So we call it the triangle,right?
Which is the persona, the roleof the person, the industry or

(14:00):
sub-industry they're in, andthen the trends on how we solve
that.
So we work with VPs ofoperations and industrial
manufacturing to solve these twothings, roughly something like
that.
Gen AI is really good at pullingout those trends.
So those are those are, youknow, a couple of big ones for
me.
Identify the top three biggestpain points.
Once you've done that, you needto have a little bit of a mental

(14:21):
shift.
And that's number two.
So once you've identified yourtop three, I want you to pick
the first one.
And number two is you have torefine the outputs to sound like
you.
You're going to need to spend alittle bit of time saying it's a
little more like this.
It's a lot more like that.
I would say it this way.

(14:41):
And again, you can create acustom GPT, you could create a
Gemini Gym or a copilot agent.
I don't really care which oneyou use.
And you're just you'refine-tuning what's what we call
the custom instructions.
And so just understand with justsome basic prompting, you can
probably get pretty far withsome of these use cases.
But step two is you have to getit to feel like you.
Why is that so important?
LLMs are really good out of thebox, like really, really good at

(15:04):
doing a lot of different things.
Another reason why I feel likepeople struggle to fully adopt
is they just it just doesn'tfeel right.
You know what I mean?
Like it's one of the like itdoesn't feel authentic or feel
quite like me.
And so we tend to, you know,over over rely on our own human
heuristics to do something,versus trying to figure out how
we, you know, get AI to help usto do it.

(15:27):
So step two is you have toinvest just a little bit of
time, 30 minutes, 45 minutes totruly make it your own.
And then the next and finalpiece.
So you've identified thebottlenecks, you've you've
started to make sure it soundslike your own, is to create your
ongoing roadmap.
And what do I mean by that?
Right.
So I I said in the beginning,let's pick three tasks.

(15:49):
The reality is I don't want youto try to solve all three at
once.
What I want you to do is pickone, train it on you, solve next
up.
And so, you know, many of youshould start to use, I don't
care if it's an Excel, I don'tcare what you know, if you use a
sauna, we use a sauna, is tocreate a roadmap for yourself to
do the next up thing.
So then I'm like, okay, thisone's in backlog, I'm gonna move

(16:12):
it to next up, and then I'mgonna deploy, right?
And then optimize in this wholecircle, right, as a part as part
of that.
And so for a lot of you outthere, if you just do those
three things, it is going tohave a massive, massive impact
on your day-to-day.
And if you're looking for asolution out of the box, I do

(16:33):
want to recommend Journey AI.
Journey AI already has a lot ofthese purpose-built agents and
assistance for account research,deal strategy, uh, even helping
with social posts.
They've got a more general onecalled Rep GPT that you can, it
can do kind of anything salesrelated.
It's like a big sales brain thatknows to help you to do
one-to-one prep or any otherrandom activity.

(16:55):
So if you want to get ahead andskip over a lot of the
day-to-day use cases, go sign upfor a free trial of Journey AI.
Link is in the show notes aswell.
Meetjourney.ai will be a bigtime saver and game changer for
you for sure.
So that's today's episode.
So, today's episode, what Ithought was really important and
what when we were working on theoutline was to really try to

(17:18):
help everyone understand wherewe're at, big picture, and then
get get tactical in terms of howdo you make sure that you're
implementing in your day-to-daybreakout use cases and not
bubble use cases?
And there's no doubt, I thinkevery organization in 2026 knows
without a doubt that we have toinvest in AI.

(17:38):
But the key thing I want you tothink about is this is if we
fast forward a year from now,are you gonna be cleaning up a
bunch of bubble-based AI usecases that never move the
needle?
Or is your organization gonna beheading into 2027 looking
completely different and workingcompletely different?
Because that is the time andplace that we are at right now.

(17:59):
There's gonna be probably threetypes of companies.
One, the company that continuesto wait and work in the old
school way.
I think we all know what's gonnahappen there.
Two, the type of company thatsays, yes, AI, and then this
group goes and buys clay, andthis group buys this, and this
group buys this.
And then in about 12 months,somebody raises their hand and
goes, guys, what are we doinghere?

(18:19):
We've invested in all thistechnology, all these things.
I'm looking at our repproductivity year over year, and
it's a blip.
5%.
What happened to these 10, 20,30 percent gains?
And then organ and thenorganization three are gonna be
the winners.
These are the organizations thatcontinue to say, okay, what are
the bottlenecks in these roles?

(18:40):
Okay, do we have a clear pathfor solving the bottlenecks for
these people?
The technology is step two orstep three.
How we're gonna solve that isstep two or step three.
Have we really identified thebottlenecks for each individual
role, what these people aredoing, the unlock to revenue or
better customer conversations?
And they continue to build anddeploy AI around that.
Those are gonna be the winners,my friends.

(19:01):
So thank you so much for tuningin.
I hope you got tremendous valueout of the episode.
And as usual, make sure if youdid get value, like the video,
subscribe to the channel, sharewith your favorite sales buddy
as well.
And if you're listening on yourfavorite podcast platform, make
sure to sign up for alerts anddownloads as well.
So I hope you got a ton of valueout of this.

(19:23):
As I talk through this, it justbecomes more clear and more
clear to me every day that thereis a right and a wrong way to
implement Gen AI.
And if you listen to today'sepisode, listen to it a few
times over, and follow the stepby step, I promise you're going
to be in that third group ofpeople that are a year from now
looking back saying, I am soglad that we took this approach

(19:43):
and didn't just start throwingevery new AI tool at the
problem.
Thanks again, everybody, fortuning in, and we'll see you on
the next one.

unknown (20:00):
Mm.