A team of 10 people on Zendesk costs $13,800 per year. With Copiloto enabled, that cost can rise to $20,000
That’s more than one of our clients' support agents earns in a year 😬
In 2023, I started building an alternative from Latin America
We don’t charge per person. Instead, we offer fixed plans based on tickets, with unlimited agents and all channels included (WhatsApp, email, Slack, Instagram)
It’s called Melonhelp, and the smallest plan starts at US$29/month not matter how many users you need
2
54
Un equipo de 10 personas en Zendesk paga US$13.800 usd al año. Con Copiloto activado, puede subir hasta US$ 20k
Más de lo que gana un agente de soporte al año 😬
Empecé a construir, el 2023, una alternativa desde Chile 🇨🇱
No cobramos por personas, sino planes fijos por tickets, agentes ilimitados y todos los canales incluidos (WhatsApp, email, slack, Instagram)
Se llama Melonhelp y el plan mas pequeño parte en US$29/ mes
El mes pasado Meta empezó a cobrar los mensajes de las empresas que dan soporte por WhatsApp
Aún hay gente que no sabe cuánto va a pagar así que hicimos un estimador en @melonhelpapp
Eliges el pais, la cantidad de mensajes, categoria y listo.
melonhelp.com/calculadora-wh…
1
30
It's really hard building something, make money and at the same time try to post something valuable on Twitter
2
24
Danilo Toro 🤖🍈 retweeted
Hola JS. El tema de garantías sobre solvencia bancaria es bien interesante, así que vale la pena extenderse un poquito.
Como introducción: Fintual es una administradora de fondos regulada por la CMF, y con el APV en Fintual inviertes en cuotas de fondos mutuos. Esas cuotas son tuyas y cada fondo tiene un patrimonio y RUT separado de la administradora. Si esta última deja de operar, las cuotas de fondos siguen siendo tuyas y la CMF puede traspasar la administración del fondo a otra entidad.
La garantía estatal de los depósitos bancarios, por otro lado, existe porque los bancos reciben plata en cuentas corrientes, cuentas vista y depósitos a plazo, y la usan para financiar préstamos. Aunque quienes recibieron esos préstamos no paguen, el banco sigue obligado a devolverte lo que depositaste. La protección bancaria estatal cubre precisamente ese riesgo de que el propio banco no pueda pagar.
En Chile, si un banco entra en liquidación, la garantía estatal protege el saldo en cuentas corrientes y cuentas vista, para personas y empresas. En depósitos a plazo, la garantía estatal cubre a personas naturales hasta 200 UF por banco y 400 UF en total (unos $8,3 millones por banco y $16,5 millones entre todos los bancos). Pero para los DAP de personas jurídicas, la cobertura estatal es 0% (a propósito, esto, junto a la liquidez diaria de los fondos, explica por qué generalmente una tesorería corporativa prefiere más invertir en un fondo mutuo tipo money market que en un DAP).
En un fondo mutuo, el valor de tus cuotas depende de cómo les vaya a las inversiones del fondo, como sus bonos o acciones, que están custodiadas en empresas externas especializadas en custodia de valores. La administradora gestiona esa cartera, pero no tiene que devolver un depósito con su propio patrimonio, como en el caso de un banco. Por eso la garantía bancaria no aplica a las cuotas de fondos mutuos, sea o no APV, e independiente de si este APV es en un fondo mutuo de una administradora bancaria o no bancaria, o dentro de un APV de una compañía de seguros que invierte en cuotas de fondos.
Hola @fintual @omarlarre . En qué categoría están uds? Su APV tiene garantía estatal? Que pasa si quiebran? Gracias!
Danilo Toro 🤖🍈 retweeted
All of us who run software consultancy thought that AI was going to kill our business
What we're seeing is companies are using AI to improve their business, not to get into software.
One of our clients is a consulting firm. They could grow their dev team and build more for their clients with Claude Code.
They decided not to. Their business is consulting, not maintaining software. When something breaks, they call us and it's fixed fast.
AI didn't kill us. It's bringing us more work.
All of us who run software consultancy thought that AI was going to kill our business
What we're seeing is companies are using AI to improve their business, not to get into software.
One of our clients is a consulting firm. They could grow their dev team and build more for their clients with Claude Code.
They decided not to. Their business is consulting, not maintaining software. When something breaks, they call us and it's fixed fast.
AI didn't kill us. It's bringing us more work.
Danilo Toro 🤖🍈 retweeted
One year ago I started building a No-Code computer vision platform named YOAO.
You upload any video and YOAO detects the objects on it and get the data inside it.
For months we were stuck at ~14% of error against human detection but now, after weeks fixing every mistake our model made, we are at 4.6% 🎉
I think I wasted my time playing Pet Society when I was 12 years old
Facebook’s games ruined my life
I'm the founder of a fintech at 12 years old, and we're raising a $1.5m pre-seed. Here's my W27 application to @ycombinator.
145
Even though this video is from 2024, the playbook hasn't changed
1. Launch quickly
2. Get your first customers
3. Talk a lot with them
4. Iterate
AI solved step 1. We have to focus on steps 2 to 4
1
30
If you want to build products companies really love, start by building for companies
That’s the playbook I followed
I found a partner who was good at sales. They sold software services while I built, and my only job was to prove we were the best at it
That is how Robotipy got to 80+ projects in 8 countries. RPA, AI, custom software.
Being inside companies is the best product research there is. You see what problems they have, and more importantly, what they are willing to pay to fix.
Support kept breaking in the same place: WhatsApp, Slack and email, nobody owning the reply. That became Melonhelp ($1.1K/m)
Traffic consultancies were still counting cars by hand, watching hours of video. That became YOAO ($0.5K/m)
Posting on X won’t teach you that
Services pay the bills and products are where you make it repeatable.
1
33
Traffic engineering is harder than it looks.
You can't A/B test a street. The only fair comparison is the same street against itself, before and after a change. And when the call is wrong, people get hurt.
So the data has to be good. Most traffic counts are still done by hand, a few peak hours at a time.
With YOAO we count every movement for the full day from one video. It's already catching mistakes the manual counts miss.
13
Shipping features is the easy part.
Talking to a few hundred users without dropping half of them is the hard part.
Most makers optimize the first and wing the second, then blame churn on the product.
16
One year ago I started building a No-Code computer vision platform named YOAO.
You upload any video and YOAO detects the objects on it and get the data inside it.
For months we were stuck at ~14% of error against human detection but now, after weeks fixing every mistake our model made, we are at 4.6% 🎉
What Anthropic is doing for the global economy is crazy
First, they launched Opus 5.5 which has almost unlimited tokens
Then they give one year of free seats to their team plan with Claude for Startups
We are coding more and better, optimizing meetings, remembering everything and new successful entrepreneurs are born
87






