Results from the field
Results from the field
Real systems, in production, with measurable results. Here are a few.
100+
work hours saved every month
Premium Insurance Network
Background
An insurance agency network specializing in holistic portfolio analysis for premium clients. Each analysis required a skilled agent to manually go through dozens of documents - policies, riders, medical approvals - and build a detailed report. One to one and a half hours per case, nearly 100 cases a month.
What we built
An end-to-end AI system that connects to email, detects incoming cases, runs a document-processing pipeline through the Claude API, and maps the data into their professional analysis format. Includes a branded, designed report page ready to send to the end client.
The result
ILS 150,000-200,000 a year in saved labor cost. A process that took 90 minutes now takes minutes. Agents focus on advisory and sales instead of reading documents. The network can double its client base without hiring.
The system paid for itself within a few months.
Stack Node.js · Claude API · Python · custom web app with auth
₪240K+
annual operational savings
National Food Importer
Background
A food import company with hundreds of B2B customers - restaurants, chains, distributors. Every order was handled manually: a phone call or WhatsApp, an employee typing the order into Priority, calculating quantities (units, cartons, pallets), generating an invoice and delivery note, and fixing errors. 15-20 minutes per order, ~350 orders a month. Over 100 work hours monthly.
What we built
A WhatsApp AI agent that takes orders directly from customers. It identifies products, calculates quantities and prices automatically, validates the minimum order, and creates the order directly in Priority ERP - including invoice and delivery note generation. We also built an AI knowledge-base agent for the whole project.
The result
A full order-entry position was eliminated - ILS 12,000-14,000 a month in salary alone. Zero typing errors (which cost thousands a month in returns and credits), accounting documents generated automatically, and customer response time dropped from minutes-hours to seconds. Total savings of ~ILS 20,000 a month.
The system saved over ILS 240,000 a year in direct operational costs and paid for itself within three months.
Stack WhatsApp AI Agent · Priority ERP API · automated document generation
