AI engineering for thefounders creating new worlds.

We build products and AI agents that connect to your data and tools to create measurable impact, solve your biggest bottlenecks and tackle your most ambitious projects

Products and solutions development that makes an impact in your organization

Customers and startups where our founder made an impact

  • Besty
  • Padelite

01 — Services

One team across product, operations and growth.

Start with the work that needs to change. We help define the problem, build a useful first version and decide what to do next.

01 — Operational agents

Agents for the work between your tools.

Build around a defined job, with the context and tools an agent needs to do it. Set limits, test failure cases and give people a clear way to review or take over.

Workflow designContext and toolsTesting and limitsHuman review
Meet the founder

02 — Embedded engineering

Build the product and the systems behind it.

Work with your team on the software customers use and the systems that keep it running. Product decisions, integrations and implementation stay in the same conversation.

Product developmentBackend and frontendData and integrationsTests and monitoring

03 — Commercial intelligence

Give sales and marketing the context to act.

Connect CRM history, company knowledge and relevant market signals. Build account briefs and source-backed drafts your team can review before anything goes out.

CRM contextAccount researchSource-backed draftsHuman approval

04 — AI adoption

Help your team choose and run its first AI workflow.

Map the work, choose a use case and practice with the people who will use it. The plan includes ownership, limits and how to judge whether it is helping.

Workflow reviewUse-case selectionHands-on practiceOperating plan

03 — What Omochi does

Connect your systems. Build the product. Put it to work.

The three steps
StepWhat happensWhat you get
01 ConnectWe plug into the tools, data and records you already run: your CRM, your ERP, your spreadsheets, your internal APIs.One place where your operation's data makes sense.
02 BuildWe build the product, agent or workflow around a real job, with the people who do it today.A working first version on your real data, with its limits stated.
03 OperateWe launch it inside your daily operation, with a person approving what matters, and run it with you until your team runs it alone.A system in production, measured against what would have happened without it.
01 Connect
We plug into the tools, data and records you already run: your CRM, your ERP, your spreadsheets, your internal APIs.
02 Build
We build the product, agent or workflow around a real job, with the people who do it today.
03 Operate
We launch it inside your daily operation, with a person approving what matters, and run it with you until your team runs it alone.

Inside your systems

Nothing lives in a new app. The work connects to the tools your team already uses.

A person signs off

Every run keeps its inputs, its approvals and who decided. Your team approves what matters.

Measured in production

We judge the work against what would have happened without it, in your numbers.

04 — Working together

How we work together.

Start by scoping the problem. Then bring us into your engineering team or give us a defined product or workflow to build. Each stage is quoted separately.

Message me on LinkedIn

AI Engineering

Build alongside your team

Product and engineering work, in the same room.

Add hands-on engineering to a clear piece of your roadmap. We work through the product decisions and technical details with the people responsible for the result.

Terms agreed before work begins

Talk through your project

What the work covers

  • Product development
  • Backend and frontend
  • Data and integrations
  • Testing
  • Review and handover planning

Product and agent builds

A defined build

Take a useful idea through to a working system.

Turn a product idea or operational workflow into a testable first version. Connect the context and tools it needs, then use the proof to decide what to expand.

Terms agreed before work begins

Talk through your project

What the work covers

  • Workflow and product design
  • Data and tool connections
  • Working proof
  • Testing and human review
  • Rollout planning

Scope

Start here

Decide what is worth building.

Look at the customer, the current workflow and the constraint. Define the first useful output, the access it needs and what would make the next step worthwhile.

Terms agreed before work begins

Talk through your project

What the work covers

  • Problem and baseline
  • Defined output
  • Data and access needs
  • Acceptance criteria
  • Next-stage proposal

Process

Scope, Proof, Scale, Operate. Each stage is scoped and quoted separately. Where you stop depends on what the proof earned, not on a roadmap.

  1. 01

    Scope

    Define the problem, the output, the owner and the data. Agree the limits, how to judge the work and what the next stage would cost.

    Set in scope
  2. 02

    Proof

    Build one working example with real inputs. Test it with the people who will use it, make its limits visible and decide whether to continue.

    Set in scope
  3. 03

    Scale

    Extend what worked to the users, integrations and operating conditions it needs. Test the risks and define where people review or approve actions.

    Set in scope
  4. 04

    Operate

    Monitor the system, learn from its use and improve it. Agree who maintains it and what support or handover the team needs.

    Support agreed in scope

05 — Founder

Deep in AI, with the startups leading their industries.

Emiliano Potignano, founder of Omochi Labs, above Manhattan

Hi, I'm Emiliano Potignano. I build with cutting-edge American AI startups like Wander, Besty, Resolve AI and IBG Legal, applying AI to every cycle of the company: product development, growth engineering and go-to-market, data science and analytics.

That work shapes how Omochi builds. Before writing code I want to know who the customer is, what they need and how the product reaches them. A useful product needs a way to find its customers; growth needs a product that delivers.

Omochi brings both into the same room. We start with the constraint, build a working first version on your real data, and learn from the people who use it before deciding what to grow.

Outside work I'm a Kyokushin Karate black belt: multiple-time Argentine champion, World Championship competitor in Japan, and the first Argentine to reach the podium at a Japanese National Championship. Discipline, method and competing at world-class level are the same things I bring to the work.

Cutting-edge brands where I've made an impact

Emiliano Potignano · Founder, Omochi Labs
AI engineeringGrowth engineering and GTMData science and analytics

Know more about me

Featured on Silver.dev, one of Argentina's most respected tech podcasts.

Growth marketing and generating demand for software companies: 55 minutes with Gabriel Benmergui.

“No hay ningún ad que pueda salvar un mal producto.”
Watch on YouTube →

06 — FAQ

Before we build.

A few practical questions about starting, scoping and running the work.

Talk through your project

Each stage is scoped and quoted separately. The proposal sets out the deliverable, assumptions and commercial terms before work begins. Continuing to the next stage is a separate decision.

07 — Start here

Build what couldn't
exist yesterday.

Tell us what you're trying to build, who it's for, or where the work gets stuck. We'll talk through the next useful step and whether Omochi is the right team for it.

  • A clear problem
  • A useful first step
  • Scope before build
  • A decision on fit