Work

The problem, what we built, and the numbers.

Selected engagements written up the way we'd want to read them: what was broken, what shipped, and what changed afterwards.

Client names are omitted where appropriate; project details reflect the work delivered.

01FMCG distribution22 weeks
A national FMCG distributor
Forty branches, six disconnected systems, and a monthly close that took eleven days.
Read the study
Days to close the month
11 → 3
Systems retired
6
Stock count accuracy
99.2%
Branches migrated
40
02Construction and contracting18 weeks
A contracting group
Project costs that only became visible after the project had already lost money.
Read the study
Cost data lag
−18 days
Field entries logged same day
94%
Active sites on the system
12
Weeks to first site live
3
03Healthcare services16 weeks
A specialty clinic network
Insurance claims rejected on formatting errors nobody caught until the money didn't arrive.
Read the study
Claim rejection rate
21% → 6%
Average time to payment
−14 days
Clinics live
8
Model outputs human-reviewed
100%
04Sales operations · Applied AILive system
AI Lead Qualifier — calling & WhatsApp
An Odoo-based qualification system that gives prospects consistent answers, then qualifies them by voice or WhatsApp before the sales team steps in.
Read the study
Admissions tracked
10
AI calls logged
14
WhatsApp chats handled
5
Qualification channels live
2
05Sales operations · Applied AILive system
AI Conversation Moderation
A review layer for AI calls and WhatsApp conversations that keeps automated customer interactions aligned with the business’s approved knowledge and qualification rules.
Read the study
Conversation channels covered
2
Shared moderation layer
1
Arabic conversations supported
AR
Review path for exceptions
Human

Your situation is different. The method isn't.

Start a diagnosis

The problem, what we built, and the numbers.

A national FMCG distributor

A contracting group

A specialty clinic network

AI Lead Qualifier — calling & WhatsApp

AI Conversation Moderation