If I Wanted to Build an AI SaaS in 2026, I’d Do This
So, building a B2B AI SaaS is probably
one of the hottest things you can do in
tech right now. But, most developers
that actually get started with this got
it completely wrong, completely in
reverse. They go into their cave, work
on something for 6 months that they
think is a great product, and then they
figure out that actually nobody wants
it. So, in this video, I'm going to show
you the path that I would recommend if I
were to start over from scratch today
and build a new software company from
the ground up. It's the path where you
actually get paid first and then get
actual proof before you call it a
product. And now, this is the exact same
methods that I've been using for the
past years already to build and launch
multiple software companies. And there's
going to be three steps to this. And
most people actually go straight to step
number three, which is where all the
problems come from. So, in this video,
I'm going to show you how you can safely
go from step one to step two and then to
step three in order to bootstrap a
software company from the ground up on
your own or with a small team without
relying on external funding, so you keep
100% of your shares. And a quick
disclaimer before we dive into things.
I'm not a software millionaire. I
haven't made the nine-figure exit or
built this gigantic unicorn, but over
the years I've built my fair share of
software. Some products making a couple
hundred dollars per month, some a couple
of thousands, but it all adds up and the
pattern that I've used throughout the
years is always exactly the same. And
now, what you should really understand
about building software is that the
hardest part of this whole business game
was never
the code. This was never the hard part.
Now, more than ever with AI coding
agents, the code is the easy part. The
whole game of software has always been
about finding a customer, and in
particular your first one, this is the
most important one, plus getting them to
pay, plus getting them, most important
part, to stay. This is the whole game of
software. Where most people go wrong,
especially engineers, technical
founders, they start with the code. So,
they start here. And then they start
building something for 6 months and
figure out they cannot find any single
customer. They're not willing to pay,
and even if they find someone to pay for
it, they don't stay. The beauty of
software is that if you get this right,
this loop of customers paying and
staying, you have what they call product
market fit. In the beginning, this is
the most important thing to work to
because if you can find customers, they
pay and they stay, you have a beautiful
business model, and that's why software
companies are so attractive. Okay, so
then how do you get to product market
fit? Well, if you don't start with the
code, what do you do, right? Well, we
flip the script and we start with the
customer. And then you end up on what I
call the surface-to-software
ladder. So, this is the model that I've
been using where you start in a very
unsexy and unscalable way to building
software, really. And that is you start
as a freelancer or as an agency, or
simply just as a developer who built
something for a particular company, and
they pay you to do the build. So,
there's no fancy sign-ups, no fancy
multi-tenant support. You simply find
one customer with a particular problem
that you can create a build for, and you
let the customer pay you for that build
so you can prove your concept first. So,
that is the overarching idea of the
surface-to-software ladder. So, now
let's get into step one to explain how
to actually get going with this. And
step number one is what I call the done
for them phase. And this is simply where
you do all the work. This is where the
surface part comes in. And the whole
goal here is to find that one particular
company that has this problem, and let's
zoom in a little bit on how we can get
started here, right? Because this is
also the hardest part to get right. So,
you might have questions, where do I
start, how do I find that first client,
what problem do I actually solve? And
the word that you want to remember and
that you want to optimize for is what I
call breakage. And breakage in a
business can be any type of
problem leakage where they are either
leaving money on the table by not
solving that particular problem or a
problem is actually costing them a lot
of money on a weekly and a monthly
basis. So, this is what you ultimately
want to look for. But the only way that
you can kind of like get to that point
is by getting insiders knowledge into
what's actually going on inside the
company. Because often, this is the very
the unsexy stuff, the stuff that goes on
in the background in certain
departments, certain employees, between
certain systems. And without insiders
knowledge, it's usually very hard to
understand what this is all about. And
most of the the big software companies,
they fix already fixed the obvious
stuff, right? So, people need to send
emails, people need to manage, people
need to plan tasks. So, you are not
going to build the next Gmail or the
next Slack or the next task manager.
Because good luck competing with those
giants, right? So, those are the that is
the obvious stuff, the obvious breakage.
Where we want to get to is is way more
niche and way more subtle. And the only
really the only way to find this out is
to actually
go and talk to people. And I know for
most engineers, this is the scary part,
right? This is This is why we hide in
our caves for 6 months and build stuff.
But trust me, that is a recipe for
disaster. You actually need to talk to
people. And instead of like cold
emailing or pitching a whole bunch of
people or going on Upwork, what I
actually recommend is where to start is
with your warm network. Pretty much
anyone that already kind of knows you
and if you send them a message, they're
not like, "Who is this person?" They at
least to some degree, there's some
familiarity there. And why you want to
start there first is because if you
compare that, for example, with cold
outreach, which is the opposite, right?
Where you reach out to someone that
pretty much, let me use some colors over
here. Cold outreach, and then we have
warm outreach with your network. The
difference pretty much is like cold,
these people have no idea who you are.
Um, for example, on a network on
LinkedIn. So,
for you to get a reply and for someone
to actually decide, "Okay, let me
actually take a Zoom call with this
person or meet in person to talk about
my business." You're going to have to go
through a lot of no's in order to
eventually get one yes. Whereas with
your warm network, it's way easier to
get an introduction and to just to just
have a chat. Because the thing is
nowadays with AI
and what AI can solve, there is breakage
in pretty much every company. Because
there's so much AI can solve right now
that wasn't possible
a couple of years ago. So, you'll find
all of these businesses that have all of
these processes like admin, support,
sales, marketing, like pretty much all
the departments of a business. What else
we got in here? We got all the ops, of
course. Operations. All these areas have
people and processes and systems
working on something. And usually, this
is in a company it like this is all very
messy. Because over the years, it builds
up. And now with AI, we can kind of like
step in. And by the way, this this nice
circle over here represents the
complexity of business. Like, I don't
know. That was just what came to my
mind. So, now with AI, we can just come
in and we can try and fix all these Let
let's see. If we have the
the different operational layers, if you
will, in a company, right? You have all
of these breakage points at different
levels of the company all relating to
these different types of departments
that things fall under. So,
we can step in and one by one we can
identify these opportunities. And the
thing is, you should position yourself
when you jump on these chats with people
not as
"Hey, I'm going to pitch my services and
I'm going to say, look, I know all about
AI. I know how to solve your business.
Here's what I can do." No, instead, you
position yourself
as curious. That's probably the best tip
that I can give you. So, instead of
being salesy, being a pitch, thinking
about you thinking that you know what
they need, you position yourself as
really curious. And you go through a
process what's in sales known as
discovery. And this is simply asking
questions. Really getting to the core
problems of what they experience within
a company. And you need to do some
digging in there. What you want to start
listening for are the following signals.
So, you want something that is kind of
like described as as painful, ideally
also repetitive,
costly, and also unfixed. If I had to
boil it down, this would be the
definition of breakage. And if you find
a process, if you find something within
the business that checks all of these
boxes, so both painful, repetitive,
costly, and unfixed, you have a serious
opportunity to make a lot of money. What
you then do is you propose a solution.
And this can come in a variety of forms,
but the key thing here is that you want
to keep it really simple, lean, and
minimal so you can build it quickly, so
you can really position this in like the
software world is what we call as a
proof proof
So, you want to build quick, and
especially nowadays with AI, this gets
easier and easier and easier to get like
in a day, 2 days, maybe 3 days, get
actually something that already works
and can show some tangible value. So,
you want to propose a POC, but there is
also one important part. You also want
to propose an actual dollar amount to
it. Because
everyone wants their problems solved if
you go out and just do it for them,
right? If you ask, "Hey, can I build
this for you to solve this problem?"
Like, "Yeah, sure. Why not? Go ahead,
try it. We'll see what comes out of it."
As soon as money is on the table, then
the situation and the dynamics switch,
right? Because now it becomes an
equation. It becomes a question of is
what we are putting into this thing, is
that worth more or less than we are
getting out of it in terms of value? And
how you can go about this is you
actually create a proposal for them. So,
here's what that could look like. You
could say, "If I build you X, and then X
is the the nice solution that you're
going to propose that solves the problem
in a very lean and minimal way, and then
given it solves Y, this can be some type
of guarantee, meaning you could say if
it solves a certain percentage of
problems, if it saves you a certain
percentage of time, or makes you a
certain amount of extra revenue on a
monthly basis." Again, it all depends on
what you are solving and what the unit
economics really of that problem are and
how you can make an ROI ROI calculation.
You can put a guarantee on that. If I
build you X, given it solves Y based on
this guarantee, would you be willing to
pay me, let's call that Z, for the
initial build plus another, let's call
this, I don't know, W
dollars per month to maintain this for
you? So, now you get to a proposal. And
the whole thing, the whole the the the
true test over here is them willing to
pay for it. This is so important because
if they're not willing to pay for it,
probably other customers are also not
willing to pay for it, and that means
the value
is not right, and that means you don't
have a product market fit yet. And then
quick thing, if you're getting value out
of this video so far, would you do me a
favor and hit that like button below
this video, and also consider
subscribing to the channel if you want
to see more content like this. That
would really help me out. Thanks. So
step number one, doing all for them is
all about the unsexy, unscalable work.
You're essentially delivering a white
cloth surface fully custom to that
particular client. And you should not
skip this, and this is also not below
you. If you look at some of the most
successful tech companies like Stripe
and Airbnb, this is exactly how they
started. Okay, now the beauty of this
whole phase one is that you actually get
paid to do this. So you don't have to
like go into this founder cheap founder
mode and like work for months on end
before you get anything. You actually
get paid to prove your idea, which is
the best thing about this approach. Now
before we get to step two, which will be
up next, there's one more test that you
need to do in order to figure out if
you're actually building the right
thing. And that is the what I call turn
it off test. Very simple.
Turn off. So if you have a great product
that a business relies on, that gives
them great value,
if you turn it off, ideally within the
next hour, you should get a phone call.
Hey, this is not working. Like what's
happening? What's going on? The sooner
that phone will ring and they will start
to call you wondering what's going on,
the better your product. So, think about
for example, a customer service or
customer support automation system. Go
this. So, this is actually something we
worked on a lot. So, customer support
automation. So, tickets would come in
into an inbox. So, there was an email
inbox. This is my best drawing of an
email inbox. And then, what would happen
is the AI would start working on this
and start processing and then would
reply to the customers, right? And it
would have an escalation gate to
escalate to a human support agent if
things go wrong.
We've had plenty of times during our
first like the first initial phase where
we were onboarding things, sometimes
something would break or like a model
wasn't available anymore. Like all of
the stuff that happens in the beginning,
right? That support team is there every
day like going through all these
tickets. And this was a huge time saver
for them. So, they only really got about
25% of the tickets that were escalated
that the AI couldn't handle. As soon as
this system was not working anymore, the
workloads of those what they had to do
went from like now they don't have to do
25, they have to do 100% of the tickets
again. So, if you do the math, that's a
4x in total work that needs to be done
with the same team. So, that is how you
should think about the
the unique economics and the ROI
calculations of software. Now, this is a
very like simple example. It's very
straightforward, but you ideally always
want to look for something like this
where you can get to these types of
calculations where if they if you turn
it off within hours, they will call you.
So, if you get to that point with your
first build, you're actually in a really
good spot to make a lot of money and it
can be very successful. I would even
make it conditional before going into
step and phase number two, where
otherwise I would just continue doing
step one over and over again until you
find something that checks all of these
boxes. Okay, so once you're there, we
can go to phase number two of the
surface to software ladder. Let's call
it step number two, and this is where
you move from fully done for them, you
move to done with them. So, this is
where you're going to do some of the
work, but you're also going to include
the client in there. And the whole goal
of this phase is to get three to five
near identical customers as your
original customer. So, the whole trick
that you're going to do here is you now
have that product skeleton, right? And
again, it's still it's POC phase. It's a
little bit sloppy. It doesn't have all
the fancy onboarding and the
multi-tenant support, but you're going
to another company that's very similar
and that also has that same breakage,
that same problem that you can solve.
And you rebuild it and integrate it
within their company again. So, you
don't even necessarily have to onboard
them to the build you already created.
You can do another like custom built.
Because in the beginning, you have to
figure out look what what is the 80%
that's the same for everyone, and what's
the 20% that is custom tailored. So,
what I recommend here is also when
you're trying to find that next
customer, up often the best starting
point is to literally ask your client
number one for other companies in their
industry. Because like people network,
people know the the industry and the
branches that they are in, right? So,
that is the best way to get your foot in
the door. And I recommend to go as niche
as possible here. So, with software you
don't want to be generic, because anyone
can build a generic task manager. Anyone
can build a generic marketing app,
right? But, for example, when we were
building this customer support
automation system, we focused on a
particular niche really. So, we did
e-commerce stores who were using Gorgias
as their ticketing system. So, this
combination was already very specific.
So, the whole goal here is to find more
customers, and then you want to slowly
work towards this, like done with them.
That's kind of like the key phrase over
here. So, you do some of the builds, but
you're also going to build some, let's
say, simple dashboards that they can
see, or onboarding steps, or certain
forms and even settings that they can
fill in to now start using the software.
So, you shift more and more towards the
customer. Now, another thing that you
want to switch here is the pricing
model. So, how to think about pricing?
Whereas in phase one, it was all about
the initial build cost, and then the
maintenance fee afterwards, on the
second and clients after that, you
ideally want to turn to some types of
value-based
pricing, and you don't have to pay as
much up front anymore on like the custom
build. So, again, for the example of the
ticketing system, we, for example, the
original build, we did like a 10K
sprint, and then I think we followed up
with another 15 grand. So, we had like
an initial 25K build, and this was
already like straight-up cash for us.
And then, with the follow-up clients,
what we were doing is we would purely
priced on tickets processed, knowing
that as people, as companies start to
process more tickets with AI, so more
tickets flow through our system, the
more value they get from our system. And
this is another software hack that's is
is always good. Whenever possible, you
always want to tie the pricing model to
the value. So, that as people use the
software more, as people get more value
from the software, the price goes up.
So, then if we zoom out and look at all
of this, step number two, your ideal
outcome should be three to five near
identical customers, and they should be
happy, they should be paying, and they
should give you testimonials. If you
have all this, then you are probably in
a better position than 99% of the people
that try to start a software company,
and you're in a great spot to continue
to really scale this thing. And this is
actually the perfect moment to talk
about today's video sponsor, Vanta,
which is a company I'm working with
right now. Because here's what almost no
one warned you about when you're trying
to sell to more companies, especially
bigger companies, is that once you have
that product market fit, they stop
asking, "Will this work?" and "How does
it work?" They start to ask like, "Is it
secure and can you prove it?" That's
when you suddenly need things like SOC 2
or ISO 27001, and whether you have them
can really make or break a deal. And
with AI baked into literally everything
right now, not only your own solution,
but also the services and tools you are
using to build and run the company, your
risk surface is expanding fast, and your
compliance needs to keep up with that.
And most people have no idea how that
works. And now, I'm going to be honest
with you, for Glide the software product
that I'm focusing on right now, none of
us on the team really started out as
compliance experts. So, that's really
why we started to use Vanta, because it
gives us one place to see our entire
security posture, meaning our internal
risks, our third-party vendor risks, and
continuous monitoring running in the
background, so this whole thing doesn't
become this once-a-year thing when you
need to prepare for an audit, but it
constantly runs in the background. It
can automate the path to frameworks like
SOC 2, ISO 27001, and HIPAA, and it also
covers the newer AI governance
frameworks everyone's starting to pay
attention to, like ISO 42001, the EU AI
Act, and the NIST AI Risk Management
Framework. Vanta plugs into your stack
with over 400 integrations and more than
16,000 companies running on it,
including ones you probably know like
Ramp and Cursor. And working with Vanta
has been a complete game-changer for us
working on Glido. And as you know, I
rarely do sponsorships on this channel
and only really promote companies that I
actually work with and would highly
recommend. So, if you're building a
software product and you want compliance
to be taken care of, then make sure to
check out the link in the description.
vanta.com/dave.
You can get a thousand dollars off if
you get started with them. So, know that
compliance will start to play a bigger
role as you start to scale. And that's
when we are ready for actually step
number three. And this is where we're
going to move into the final stage. And
this is what I call the done by them.
So, remember, right? It's you do it for
them, you do it with them together, and
this is where you move done by them,
where you actually turn the product into
a software product with self-onboarding,
self-checkout. Pretty much the whole
goal here is that you work towards a
solution is where your customer can get
results without you being in the
picture. So, while you are out there on
a beach somewhere, wherever you want to
be on a beach, like there should be
someone that can go to a website, can
create an account, can sign up, and then
ideally get results as quickly as
possible. So, this and only at this
stage
are you actually going to spend more
time on usually the engineering apart
engineering part around things like the
onboarding and multi-tenant support. And
this is things where this is also the
stage where things like your funnel and
plus website are going to be more
important because you need to draw
people in. It needs to make sense.
People need to understand it. They need
to be able to log in with an email, but
they also like the sign up with Google
button, right? So, all this complexity
grows. And trust me, there is like going
from a done with them setup where you
have 80% of the build ready and you do
the last part for them going to like
fully done by them, most people
underestimate this. Because
even with what you can do with coding
agents nowadays, yes, you can move
really fast, you can build the MVP, but
everything that comes around it, things
like
emails on the onboarding, password reset
flows, creating the accounts, magic
links, two-factor authentication, sign
up, follow-up emails, all this
complexity around it, this is where a
lot of people actually get stuck and
underestimate what it really takes to
run and scale a software company. So,
that's why this is step number three,
done by them. So, really what you should
optimize for in this stage, you're
you're already making money, you already
have a couple of customers, how fast can
you give them that result? Because
really, most products, probably say 90%
of products lose there. So, what happens
within like the first 5 minutes that you
onboard to that application? Can people
figure it out themselves and start to
get a quick win or value from it? How
easy is it? And most people will simply
not stick with it beyond the trial phase
or the first month and they will churn
and you lose them forever. So, the
faster you can get them this win, the
better. Now, then, the next big thing
that you want to focus on is
distribution. And this pretty much means
how are you going to get your product in
front of more people. Now, this is not
going to be a video on how to market
your software business. There are like
so many ways to go about this, but I'm
going to give you one tip. Most of
successful products, they don't just
rely on like paid ads because this is
like this this is the easy option,
right? But it's also very costly usually
in the beginning to get this profitable.
So, the best products have a built-in
distribution mechanisms like referrals,
sharing it with people, getting some
types of bonuses. If you look at the
early days of Facebook, for example,
this is how they grew big because a
social media platform is only as good as
the amount of people that are on there,
right? So, they incentivize to invite
people on there and there's this network
effect. So, that is one thing that you
can try and optimize for if you can. A
network effect, another recent great
example is school platform to host
community and courses. On school, you
have a 40%
lifetime
commission
on every referral or sale you make. So,
if people sign up using your link, right
now the plan was always kind of like 99
bucks a month, you would get lifetime
40% off of that. So, that's a built-in
incentive. Not only has it does it have
a network effect because you can invite
people in the community, there's also a
motivation for people to start promoting
it because this is actual actual money
if you build that up. So, that's the
goal of phase number three. You're going
to move from doing some things with them
to people being able to fully onboard
them themselves and then also getting
value quickly. That is what you should
optimize for. Then on top to grow, focus
on a distribution mechanisms. A network
effect, a built-in distribution
mechanism is always the best and then
you can expand to other forms of
marketing. So, that concludes the
surface to software ladder. You start
level one, done for them, level two,
done with them, and then level three,
done by them. Now, the one thing to keep
in mind is that you can jump off or stay
at any of these levels. Not everyone is
going to make it to level three. Let me
actually rephrase that. Almost no one is
going to make it to level three by
definition because it's really hard to
build a sustainable growing software
business. And now the great thing is
that even at level one and level two,
that's a great business and you can make
great money there. Now, you may not get
to the multi-million dollar exit, but
let's be real. The chances of that are
very low anyway. And by following the
exact process that I've explained over
here, you can greatly reduce the risk.
You don't need external funding. You can
actually get paid to prove your idea and
have fun in the process without
disappearing for months, not getting
paid, and then figuring out that it
wasn't worth it at all. And that's why I
love this model and will continue to use
it as well with my software development
company. This is the exact process that
we continue to do over and over again to
spot opportunities and to try to turn it
into a product. And with that, we have
come to the end of this video. Now, I
know that most of the people watching
this video right now will kind of like
click off and think like, "Ah, maybe
someday." But I know that a small
percentage of you watching really wants
to take this seriously. And in order to
get started with step one, doing it for
them, one of my favorite ways to really
do this is to first position yourself as
a freelancer and get started with that.
So, I have an entire video here on
YouTube on how I would start over as a
freelancer, which is something I've been
doing for years already. I also help
other developers and data professionals
to get started with that. So, if you
want to know how to find your first
project as a potential opportunity to
start working on this, make sure to
check out this video next. It's here on
YouTube completely for free and I will
share everything that you need to know.
Continue with YouTLDR
Analyze another video with Pro
Process a new video, search every timestamp, compare sources, and keep the result in your library.
More transcripts
Explore other videos transcribed with YouTLDR.

ผมสัมภาษณ์มา 2,000+ คน: กับ 5 บทเรียนที่คนเก่ง “ไม่ผ่านสัมภาษณ์” งาน เงินเดือนหลักแสน
Hunter B · Thai

ล้มเป็นพันครั้ง แต่ไม่ยอมแพ้ กับคุณกระทิง พูนผล | Mission To The Moon EP.2779
Mission To The Moon · Thai

2º LEILÃO VIRTUAL RAÍZES NELORE JACAMIN | 15/11/2026 – 14H
Canal Terraviva · Portuguese (Portugal, Brazil)

I built this app on the side for fun. It just hit $100K
Starter Story · English

Day 10: 14 Days of Prayer & Fasting | Opening The EAST GATE- The Gate Of The Coming King .
Francis Myles International · Spanish

Day 9: 14 Days of Prayer & Fasting | Opening The HORSE GATE - The Gate Of Spiritual Warfare .
Francis Myles International · Spanish

Day 8: 14 Days of Prayer & Fasting | Opening The WATER GATE - The Gate Of Divine Revelation.
Francis Myles International · English

Day 7: 14 Days of Prayer & Fasting | Opening The FOUNTAIN GATE - The Gate Of The Holy Spirit .
Francis Myles International · Spanish

Day 6: 14 Days of Prayer & Fasting | Opening THE DUNG GATE - The Gate Of Repentance.
Francis Myles International · English

Day 5: 14 Days of Prayer & Fasting | Opening The Valley GATE - The Gate Of Brokenness .
Francis Myles International · Spanish

Day 4: 14 Days of Prayer & Fasting | Opening The OLD GATE - Returning to Ancient Paths.
Francis Myles International · Spanish

Day 3: 14 Days of Prayer & Fasting | Opening The FISH GATE - The Gate of Evangelism.
Francis Myles International · Spanish