Case Stories

How Choco automated 8.8M messy orders without changing client behavior

Choco stopped trying to force LatAm chefs to use B2B portals. By using OpenAI to ingest chaotic WhatsApp voice notes and blurry photos, they doubled sales productivity.

KytoAI & Automation Firm
·
May 1, 2026
0

Key Takeaways

  • 1LatAm B2B runs on WhatsApp voice notes, not clean ordering portals.
  • 2OrderAgent uses OpenAI to map vague terms like 'the usual' to specific historical SKUs.
  • 3Automating the mess doubled sales productivity and eliminated up to 4 hours of daily manual data entry.

[@portabletext/react] Unknown block type "tableBlock", specify a component for it in the `components.types` prop

Try this translation test today

Take your 50 messiest WhatsApp orders from yesterday. Paste the transcripts into ChatGPT. Prompt it: 'Map these unstructured orders to this list of standard SKUs and output a clean CSV.' You will see the translation layer work before writing a single line of code.

Preguntas Frecuentes

How does AI translate a vague order like 'the usual'?

The AI cross-references the transcript against the specific client's order history, identifies the exact product they buy regularly, and matches it to the supplier's SKU.

Did Choco force clients to stop using WhatsApp?

No. The breakthrough happened when they stopped fighting client habits and built an AI ingestion engine specifically to handle raw WhatsApp audio and text.

Case StoryOpenAIWhatsAppLogisticsAutomation
Compartir artículo

Kyto

AI & Automation Firm

We design and build AI automations and business operating systems. Agency results + Academy sovereignty.

¿Listo para automatizar?

Construyamos Tu Sistema Operativo.

Reserva una demo gratis para ver cómo la automatización con IA puede transformar tus operaciones.

Reservar Demo Gratis