Andres Fonseca, Founder & CEO of ITRMachines talks about why they are moving away from their existing B2B GTM as they look to grow further.

• How ITRMachines help institutions by helping with their machine learning model needs (like forecasting, predictions etc.)
• How do their 20 customers pay them around $10k MRR
• How recommendations from existing customers work as a primary acquisition channel for them
• Why having to go through a long model co-creation process with customers is not letting them scale easily
• Why do they plan to pivot & serve a different market ( due to the current sales process )
• Team, founding story & external funding details

You can also watch the video on youtube here.

this is an amazing business model, right?

But the only sort of thing that I see here, right, is that your price point

is a bit too low per model, obviously.

Right?

So either you've gotta build more models for a company and sort of scale that

account to let's a hundred K, right?

That's one way of growing all.

You've gotta figure out a different strategy because I

mean, eight months for a $5,000.

Might not make sense for you in the long run.

Right,

for sure.

Exactly.

That's why we are jumping to, to the other business model.

That's, that's the key reason.

hello everyone.

Welcome to the B2B SAS podcast.

Today we have and Andreas Fonseca with us.

Andres here runs a company called Intelligent Trading Machines.

Hey Andres, welcome to the show.

Hey, you Benter.

How are you man?

I'm doing pretty good.

So, Andres, let's talk about your company and product first, right?

So what does intelligent trading machines to end?

Why do customers pay you

money?

Okay, thank you.

So let me have like a quick intro.

Uh, intelligent Trade Machines is basically a SaaS platform,

which is focused on, uh, deploy algorithmic trading across Latin.

Um, and we also have a focus on artificial intelligence, um, as a service.

So what we create is models for predictions for forecast, and,

uh, we deploy this solution.

Institutional, let's say banks, fiduciaries, corporations

that want to have artificial intelligence in their companies.

Uh, we offer that product as a service with a platform, uh, and basically

the clients consume this, these, these, um, forecast through a p I.

Okay.

Uh,

so let's talk a bit about your, uh, customers, right?

So who exactly are you serving, and is this primarily a SaaS model or do you

also have any services layer on top of it?

Perfect.

So

let's say like this, uh, we first

focus on a B2B two solution.

So business to business solution in which our clients are, are basically

institutions, uh, here in Latin America, let's say bank broker experience,

uh, pension funds that want to use, uh, both of, of these products.

Um, nowadays.

So like our core solution is artificial intelligence as a service

being like our core, core service.

Um, and yes, basically our clients consume directly as a SaaS platform, uh, which is

half like a mix of, of, of the solution because at the first stage we have like a

consistency process in which we co-create with the client, the, the AI model.

And after we create, or we co-create this AI model, so then we provide

this platform in order that the client can consume the forecast, the, like,

the outcome of the, the output of the, of the artificial intelligence

model, uh, through the platform.

All

right.

So how many paying customers do you have on your platform as.

Okay.

Nowadays we have around 15 to 20, let's say 20.

Let's say 20.

And uh, like, can you, uh, give me how much approximate

revenue did you do last month?

Approximate numbers?

Total.

Okay,

give me a sec.

It's around, um, 10 K.

Basically we, for, for every client, it's approximately like 500 K per,

uh, for the, for the platform.

Like the, the, like, the constant, uh, let's say like this payment

that we receive for, for the service, it's around 500.

Uh, I don't count, for instance, when we create a specific model.

That, like the business model is different and it's a consultancy, so

that is not like a recurring payment.

Mm-hmm.

. So that's why I, I just take out from, from, from my, like, from my map.

But yes, basically, just to clarify,

it's 10,000, uh, monthly recurring revenue across

all of these 20 odd customers.

Right?

Exactly.

Exactly.

Okay.

And okay.

Uh, so, uh, help me understand where you got these customers.

I wanna understand how you sort of, uh, found these customers in the first place

strictly from a top of funnel perspective.

Yes.

Okay.

Okay.

So basically we start intelligent Trade matchings approximately,

uh, uh, five years ago.

Um, and, uh, like two of the co-founders were part of the industry.

Like, uh, they were working in financial institutions.

So basically our first approach was like, go and knock the door for, for

our friends , uh, in the, in, in, in.

Let, let's say, um, corporations, financial corporations and show

what we were doing and like, let's say the organic growth for based

on, on that is like basically, um, like the, the just recommendations,

you know, that is like how we can, how we have achieved most of our.

Okay,

Yeah, so

just wanna understand as of today, right?

So if you get a new customer, so I'm guessing you gotta develop

a bunch of custom models for, for that particular customer.

So how much time does it take for you, uh, to sort of build the cus

So to talk, to, interact with that customer and build that solution?

How much time and effort does it take for

you

as let, let's say that's, that's the

reason why we have made the pivot.

But, uh, but yes, like, you know, when, when it is a B2B model, patient

consultancy service, uh, let's.

Just in terms of the negotiations, it take around two to six months,

uh, just in negotiation phase.

Mm-hmm.

, uh, after we finish fi, finalize the, the negotiation, then we jump to,

let's say, construct or co-create the, the, the AI model, which depend on

the problem that, that the end Sure.

Client have, uh, it'll, it'll come from two to four.

Creating the AI model.

Sure.

And once edits create, it's just deployed across our platform, which is like,

let's say, uh, eight day a Labor days.

And that's, that's like, uh, all the, the Times associate

from.

Okay.

Okay.

So, so I wanna understand, so when exactly during this stage, does a customer

pay you that first dollar of revenue?

Is it after that six months negotiation period?

Uh, so let's say yes, after the six months of negotiation.

Yes.

Okay.

Uh, so, uh, I'm just trying to do some basic calculations.

You mentioned you got it on 20 customers and you got it around 10,000.

Mr.

R that means your annual contract size per customer is around $5,000 approximately.

Right?

So can I assume that

like hundred 20.

120?

Yes.

120, uh, thousand dollars per.

Sorry, I got confused.

How is that $120,000.

So was that 10?

10?

So was that 10 k, $10,000 per customer?

No, no, no, no.

So $10,000 per month across all the customers, right?

For for the 20 customers?

Yeah.

Yeah, yeah.

So I meant, I meant I won't, I won't.

Like how much does the single customer pay you in a year?

Okay.

Okay.

That's around $5,000,

right?

Yes.

Yes.

Like six, $6,000.

Let's call it $6,000, right?

So I think so.

Uh, My question is, uh, so how do you sort of expand the $6,000 revenue

right, to something, let's say 60,000?

How exactly are you expanding them?

Sure.

Uh, so basically we expand creating more AI models.

Sure.

So that is like, uh, the nature.

So let's say this, these calculations were only based on the.

But at the end, one, all the other source of revenue is also like the consultancy

process, which basically is where we.

Much of the money, but yes.

So,

uh, I'm just trying to understand, right.

So why would you make more money during doing consultancy?

So are you consulting that company for some other help as well, apart

from just selling this or are you

No.

So we, we are consultancy the companies or in a co-creation

stage for create the AI model.

Yeah.

Because that is like, let's say like the most important part.

When, when you create an AI model, like, so that's why most of the

revenue, uh, for, for every AI model comes from, from, from the consultants.

So help me quantity consultancy.

Like

the co-creation.

Yeah.

Got it.

So help me quantify this, for example, for a $6,000 customer

value, customer ticket, right?

Ticket size.

Yes.

How much do you make, uh, during that consultancy phase?

Yeah, it's, it's interesting because let's say,

uh, uh, AI model could cost, let's say like the creation of AI model could

cost from five to 15 or even 20 k.

Mm-hmm.

, like there is where you make the money at the end of the day.

Okay.

No, uh, I'm trying to understand why not make that part free and why not

just charge your customer on a record recurring basis forever in the future?

Why exactly are

you in charging?

Um, yes.

Yes.

Okay.

Okay.

Got it.

Got it.

Got, yeah, good point.

So, uh, basically, um, we, we don't make that part for free because at the end it's

the way that also at the other side, the client is aligned in the same incentive.

If the client, if for the client doesn't.

Basically they don't have the same incentive to co-create

at the same level, you know?

So, okay.

That is one of the first case

you want.

You wanna pick serious people that that's fair.

Okay.

Exactly.

Exactly, exactly.

So I don't want to Exactly.

You got my

point.

Exactly.

Yeah.

So, no, no, my question is, right.

So that's okay.

Right.

So you've got, you've earned 1520 K during that consultancy phase.

That's obviously a cherry on top of it, right?

But eventually you gotta make money out of this SaaS revenue, right?

So you.

Get money out out of this monthly recurring payments.

Uh, so, so what exact, so how exactly are you gonna expand this?

I'm assuming like Sure.

Yeah.

Sure, sure, sure.

Got it.

So basically what you just mentioned is what we also have noticed, like if

we want to scale our product like we have like, A restriction or a block in

which we can scale our growth at Fast.

Fast.

So based on that, what we have thought or what, we are already working on

that and we are basically working with Ocean Protocol we see, which is like

the, um, Bloomberg for blockchain.

Eh?

So what we are creating is sub.

In which every data scientist can go create its own like type.

Its uh, own code, create its own AI model, deploy and sell it through our platform.

So the data scientists across globe with leverage on our technology.

To create, deploy.

And also at the end of the, they charge for their creation.

No, no, no.

So I was talking about, yeah, I, I got your future plans.

That, that makes a lot of sense.

You wanna democratize this?

My question is, you've got existing customers who are paying, you

own five, $6,000 a year, right?

Yes.

So, so my, uh, so is there a way that you could sort of

increase that, uh, amount, right?

So can you ask them to pay you $60,000?

Is that even possible?

Okay.

Uh, for the SaaS platform, no, it is not possible.

For, for, for the specific SaaS platform is not way How possible.

How by exactly.

By how can increase the amount that we, I can earn from a client.

Mm-hmm.

, let's say, eh, increase in the amount of.

Models mm-hmm.

that he will create.

Sure.

So that is my other way.

So

not only but that, that would still mean you'll

have to create more models.

And that's, that, that's like a loop beginning again.

Right.

Exactly.

So, but, but don't you have price ti uh, price, you know, along the access

of number of requests served per model?

Uh,

actually we were thinking about that, but

that for, because of my business model right now, which is b2.

Generates a lot of friction.

So actually like our, when we start, we were with uh, uh, um, API calls, like

from, from different, uh, trenches.

Mm-hmm.

. And based on that, we charge X amount or another amount.

But at the end, uh, like.

B2b.

They, they don't like that type of, uh, like charging modem.

So basically we just jump to a standard fee and, and that's it.

Okay.

But yes, at the end, for instance, for, for, for the, for, for this, um,

Real estate, real beer estate fund, we are charge or we're, uh, thinking

to just charge for the a p I call.

Sure, yeah.

Because at the end, the service, it's finally consumed from the

client of my clients, you know?

Mm-hmm.

. So it's like, that is the way we are looking.

But yes, like right now, we don't, uh, charge more for

the clients for every call.

Let's, let's talk about the growth a bit, right?

So 12 months before now, where were you at in terms of monthly recording revenue?

I think we were, let's say, I think this year

we have increased almost twice.

So I think we're half of the revenue we were perceiving.

It was half.

Yes.

Okay.

So, and most of the new customers that you're getting, most of them are

from, you mentioned recommendations.

Recommendations.

Of recommendations, yes.

Okay.

And have you tried any marketing channels?

Have you tried to expanding beyond?

Existing sort of closed?

Yes, actually we have

like, let's say a marketing strategies based on newsletters.

Also some co podcast.

But at the end you also know like the decision makers, uh, they are not,

uh, users that consume this type.

And based on this, They will buy these type of solutions.

No, there are more.

No, no, no.

Yeah.

But yes, but

yes, I totally get it.

I mean, this is an amazing business model, right?

But the only sort of thing that I see here, right, is that your price point

is a bit too low per model, obviously.

Right?

So either you've gotta build more models for a company and sort of scale that

account to let's a hundred K, right?

That's one way of growing all.

You've gotta figure out a different strategy because I

mean, eight months for a $5,000.

Might not make sense for you in the long run.

Right,

for sure.

Exactly.

That's why we are jumping to, to the other business model.

That's, that's the key reason.

But, but I, I think it's quite interesting because at the end of the day, we want to

empower data scientists across the globe.

Uh, no matter where you.

No matter your conditions, if you have like the right

information, you can create your

model, your, so, yeah, I, I got what you wanna do.

So help me explain how, how's that been working for you so far?

Have you onboarded any data scientists?

Are independent creators sort of building models on your platform?

Are they able to sell?

Are there any initial.

talk about those.

Sure, sure, sure.

Uh, so basically we have, um, worked with Ocean Protocol.

Actually we have, uh, so far achieved three grants from them.

Mm-hmm.

, like, uh, for, for like start creating elements, uh, through or for, for

this vision for democratize ai.

The globe.

Um, and we have been working with Ocean Protocol, uh, on this.

Uh, and so far we have let, let's say like almost.

10,000 requests for our libraries, like our AI libraries that we have create

to them, uh, on the ocean protocol.

So that is one of the key factors that like,

uh, so, so what protocol, uh, sorry, I don't know what it is.

What?

Ocean Protocol.

Okay.

Okay.

So Ocean Protocol is like the Bloomberg for blockchain.

Mm-hmm.

. So, um, you can buy, uh, Through tokens, and it's basically

based on decentralized data.

Okay.

So, okay, got it.

So you mentioned you've got like 10,000 users expressing interest for

your, uh, a API essay, sorry, your

platform?

For, for, for our libraries.

Just for our libraries.

So now with the, with the last grant that we achieved with Ocean Protocol,

what we are willing is to create at the app, um, for, for for a scale the.

That is, that is like, I think in two months it'll be almost done at the app

and we will have, uh, hopefully better understand of the market and understand

how at the end of the day, uh, let's say, also like from, from the part

of the creators, which are like the data scientists and from the, the,

the part of, of the consumers, which will be basically any corporation that

want or to create AI models mm-hmm.

So we will understand how, how much are willing to pay.

The consumers, and which is the best way to recognize the, the IP rights

for, for our, um, data scientists and also how we can sustainability co.

How we can give sustainability to these, to these creators.

Got it.

And what, what are you gonna do with your existing business?

Are you gonna kill it or are you gonna kill

No, no, no, no, no.

At the end, at the end, let's say like this.

Our existing business model give us, um, recognition, but not only recognition,

it's also like a label of quality.

Mm-hmm.

. Uh, so that's basically what we, we are still maintaining because

uh, yes, we need also labels

So yes, that's, that's, that's how it

is.

Alright, so let's, let's wrap this up.

Right.

When did you start the company?

Uh, we start almost five years ago.

Yes.

Like a, as a startup without funding , which was another big mistake.

I won't say it's a mistake, but that's still okay.

Okay.

and how many founders do you have?

Uh, uh, we start four, two, like

computer science guys and two

economists.

Okay.

And do you have any other folks in your team as of.

Yes.

Right now.

Right, right now the company has really increased.

Uh, as I told you, like right now, the, like, AI is one of the core products,

but we also have much more products.

Mm-hmm.

, uh, actually we have create like, um, um, the centralized wallet.

For, for interactions across latam, also based on all our knowledge.

So that is also another source of income, which is like, uh, um, much more

scalable and much more like massive.

Um, and, uh, we have other products that led us as scale.

Mm-hmm.

, uh, during, during this last year.

And, and yes, like at the end we have to manage d different sections of our

company.

And how many, how many total folks in your

team right now?

Let's say four.

And.

Like almost 20, 20 guys.

20 right now.

Yes.

Working.

Right.

So looks like day

by day and you're completely bootstrapped, right?

You're using your own money

to run the company.

Exactly, exactly.

Right.

Now, actually with, with the, let's say the centralized wallet, uh, we

are now in fundraising with that.

Uh, because we have for, for, we have, let's say, scale the product we have

almost, um, also like, uh, robots based on our technology, the product.

But at the end, this will need like it's own entity apart because.

I am pretty sure that could be like by its own like a unicorn, so Yeah.

Yes.

So, so, got it.

So like what exactly do you call your company, right?

Is it a SaaS company?

Is it an umbrella of umbrella of companies?

Right.

Is it a consultancy?

It's hard.

It's hard because we

have a, like a lot of, uh, business models.

But I, I, I, I like to define intelligent trading machines as a enabler.

And

you've got a pretty interesting name as well.

Exactly.

Exactly.

Uh, which is basically, um, focused on like a, let's say

a software developer house.

Focus, for instance, in ai, like different niche, but at the end we leverage on

the whole, eh, technology across, uh, the entire, um, ecosystem of products.

Got it.

Alright, Andres, thanks for taking the time to talk to me.