Teams from six companies at the end of La Mesa, the multi-company edition of Mate & Build

Give your team a day to build what they've been putting off.

Your team works on the company's real problems, with expert AI mentors by their side, and ends the day with something functional and scalable.

// builders, not consultants

Teams from PedidosYa, Mercado Libre, Despegar and Calm have already built with us.

  • Cashea
  • Calm
  • Silo Real
  • PedidosYa
  • Skipit
  • deRentas
  • Fondos Fima
  • Feedback Spread
  • 1.8 Music
  • Fardo
  • TurismoCity
  • Despegar
  • Mercado Libre
  • Visible
  • Rufus

How we work

  • Half Day · one team
  • Full Day · several teams

In-company buildathon

One day where your teams build with AI on their own processes. In the afternoon there are live demos and everyone takes home what they built.

What a buildathon looks like →
  • Company brain
  • Growth and operations with AI inside

The team that builds with you

After the day, we join your team and build what comes next with your people, week by week.

How we join in →

What is a Buildathon?

Hit pause. Give your team
a day to build what they've been putting off.

Half DayOne team, half a day, one problem in depth.

Full DaySeveral teams in parallel, a full day, demos at the end.

  • The Calm team watching the demos at Calm Builds

    Full Day · Friday, July 3, 2026

    Calm

    Seven areas of the same company building at the same time, with live demos at the end.

    30
    challenges the teams proposed before the day
    15
    solutions working the same day
    +20 hrs
    per week no longer lost to repetitive tasks
    7
    areas of one company, building at the same time
  • The multi-company Buildathon room, with the teams at work

    One morning · Saturday, June 20, 2026

    Multi-company Buildathon

    Six teams from different companies, in the same room, turned problems from their work into working flows in one morning.

    6
    teams in the same room
    24
    people building at the same time
    6
    builds published, with what they do and how
    1
    morning, from idea to working flow
  • Full Day · Friday, October 9, 2026

    Fardo

    Agents for sales reports, automated invoicing and digging into metric drops.

    14
    people, the whole company
    9
    processes prioritized before the day
    3
    agent flows: sales, invoicing and metrics
    +7 hrs
    building, from 10:00 to 17:15

The team that builds with you.

You have more opportunities than hands to execute them. We join your team and build them together.

Three fronts. Start with the most urgent one.

Teams

Same team.
More capacity with agents.

We raise the AI level team by team, until they build on their own.

If your team is already stretched thin and growing means hiring.

Growth and operations

From idea to something running, in weeks

We define which processes change, which growth opportunities open up and what we build first.

If the opportunity is now and the roadmap is already full.

Company brain

What your company knows, one question away

You ask the AI you already use and it answers with context, telling you where it got it from.

If you've grown and now there are tools everywhere, manual processes and questions nobody can answer.

So growing doesn't depend only on how many people you can hire.

See how we work month to month →

What we've built.

A build is something we make and leave working, published with what it does and how.

Buildathons

  • The Mercado Libre team with their buildThe seller dashboard, with sample data

    Mercado Libre

    • Operations
    • Commercial

    Seller status dashboard

    Building each seller's status by hand, every monthYou ask in plain language and get a dashboard that says where to push

  • The TurismoCity and Despegar team with their buildThe daily ad spend report, with sample data

    TurismoCity × Despegar

    • Marketing
    • Performance

    Google Ads report in Slack

    Checking campaign spreadsheets by handThe report lands in Slack, summarized, every morning

  • The Skipit team with their buildThe onboarding agent's flow, with sample data

    Skipit

    • Operations
    • Product

    Merchant onboarding agent

    Setting up every new merchant by handAn agent does the setup on its own, inside Skipit Pay

  • The PeYa Wallet team with their buildThe competitive radar, with sample data

    PedidosYa · PeYa Wallet

    • Strategy
    • Product

    Fintech competitive intelligence

    Reading every piece of competitor newsOnly what matters arrives, already analyzed

  • The UdeSA team with their prospectorThe prospecting flow, with sample data

    Team UdeSA · Corte

    • Commercial
    • Sales

    Prospector that writes the first message

    Looking for new clients by handFinds the company and the decision maker, and writes the message

  • The UdeSA team with their industry radarThe market research, with sample data

    Team UdeSA

    • Strategy
    • Product

    Industry Radar

    Twenty open tabs to understand a marketTop 5, insights, actions and sources, by email

  • Calm's leadership team at their first Buildathon

    Calm

    • All areas
    • One day

    +15 solutions working in one day

    Calm's first Buildathon: +30 challenges proposed by the teams, +15 solutions working, +20 hours a week freed from operational tasks.

  • Pato explaining in front of a screen during the Fardo Buildathon

    Fardo

    • All areas
    • One day

    Sales reports, invoicing and metrics, with agents, in one day

    A 14-person company brought its problems and left with three agents running.

1:1 sessions

Clients

  • A frame from the interface engine: the competitive radar with sample data
    • Product
    • Design

    The interface engine

    The scenes on this page: the request, the work and the product running, built with the pieces of the Design System.

To do something serious with AI, your company has four paths.

  1. A training

    Theory and other people's cases.

    Leaves judgment, not lasting capability.
  2. A consultancy

    A diagnosis and a plan.

    Expensive, slow, and leaves little in your people.
  3. An in-house team

    Hiring AI profiles.

    Takes months to put together.
  4. Mate & BuildA team that builds with you

    We don't hand you an AI roadmap to execute: we execute it with you.

    Lasting capability, without the fixed cost.
Pato Iturraspe presenting at a Mate & Build event

Pato IturraspeFounder · Mate & Build 🧉

Who's behind it.

Mate & Build is a team of senior builders, and in every company we train the people who keep building inside. That's what lets us run a buildathon with every area at once and stay afterwards.

Mate & Build was founded by Pato Iturraspe, who comes from marketing and growth at LATAM unicorns and builds with AI every week, in public. Pato has always been the one who understands the problem before the solution, and with AI that turns out to be exactly what's needed.

Every week Pato documents the builds on @patoiturraspe_ and on LinkedIn: AI experiments, mistakes and the systems that come out of them.

Frequently asked questions

Do I need to know how to code?

No. Teams build with AI on their own work, with someone by their side. What you need is a real problem that eats time every week.

Where should we start?

Almost always with a buildathon: in one day you see what your people can build and where the value is. If you want to go slower, Arranque is two sessions so the team starts using AI with judgment. We figure it out together on a 15-minute call.

What stays when you leave?

The team that learned to build, every flow written in five steps, and the access in your name. If it can't be explained in five steps, it doesn't get automated.

How long until something is working?

At a buildathon, the same day: every team closes the afternoon with its demo working. Working month to month, we first define the scope and the quick wins: within the first weeks something is already running.

What happens to my company's data?

We build on secure services like AWS, Supabase and Vercel, and the data stays protected with per-user access. Control always stays on your side: you keep using what we built after we're done.

Shall we build it at your company?

A 15-minute call to see where you stand and where to start.

Pato replies, not a bot.