01
Private Rental Owner Portal - Côte d'Ivoire
A French, FCFA-ready rental operations portal combining Telegram intake, owner-managed bookings and expenses, and current performance reporting.
Private client portal
AI Workflow & Automation Portfolio
Practical automation systems for business operations, AI workflows, and underserved communities.
AI Workflow & Automation Project Manager
Senior project management background applied to practical AI automation, operational systems, and accessible digital services.
Portfolio Overview
Practical AI automation for operations and accessible services.
Drawing on senior project management experience, I design and test practical AI-assisted workflows and operational tools for teams, small businesses, and field-based services.
For businesses, I design tools that reduce manual work, structure operational data, improve coordination, and turn everyday messages into trackable actions.
For communities and field teams, I focus on accessible interfaces like SMS, Telegram, and spreadsheets so AI-powered services can work where complex apps are not practical.
Featured Workflows
02
Property Management Automation Tool
A tested service fee workflow for one property, using Google Sheets for updates, Telegram for manager control, and Supabase for payment and reminder records.
Working system
03
Farmer AI Assistant
An SMS-based assistant that connects farmers to weather information, supplier data, crop guidance, and automated alerts.
Published n8n templateTest mode
04
Asana AI Project Manager
An AI-assisted project management workflow that turns Telegram messages into structured Asana tasks and task updates.
Working demoProject management
Featured Workflow 04 · Working Demo
Asana AI Project Manager
AI-assisted project management workflow that turns Telegram messages into structured Asana tasks and task updates.
Working demoProject management
How It Works
A user sends a task request through Telegram. AI is used to interpret the message, identify the task action, structure the task details, and route the action into Asana. Follow-up messages can search for an existing task and mark it complete.
Why It Matters
Project teams often lose time turning informal messages into structured tasks. Important requests stay buried in chat, task details may be incomplete, and project managers need to manually organize follow-ups inside project management tools.
Result / Output
- Telegram messages become structured Asana tasks.
- Existing tasks can be found by name and marked complete.
- The user receives a confirmation once the action is done.
Operational Impact
Closes the gap between informal communication and structured task tracking, making it easier for project teams to confirm progress at any time, from anywhere.
Design Logic
Telegram captures the request, AI turns it into a clear task, and Asana stores the task for follow-up.
Featured Workflow 03 · Working MVP · Test Mode
Farmer AI Assistant
SMS-based AI workflow for farmer support and automated alerts.
Published n8n templateTest mode
How It Works
The workflow receives an SMS, checks the farmer profile, detects the request, generates a practical reply, and logs the interaction for follow-up. AI is used to classify unclear SMS requests and support intent detection; farmer location, crop, language, opt-in status, and weather zone data guide the reply.
Why It Matters
Many farmers in developing countries need timely information, but complex apps are not always the right interface. SMS remains more accessible in many operating environments, especially for users who need simple and fast responses.
Result / Output
- The workflow prepares and logs practical SMS replies based on the farmer's request. Outbound sending can be enabled for controlled delivery tests.
- Messages, responses, farmer needs, and alerts are logged for follow-up.
- Field teams get structured records instead of scattered communication.
Operational Impact
Helps cooperatives serve more farmers through a basic phone channel while improving visibility on recurring needs, location-specific risks, response history, and fire or flooding alerts.
Design Logic
SMS reduces access barriers, Google Sheets keeps the database easy to maintain, weather data supports everyday decisions, and AI helps classify unclear farmer requests.
Featured Workflow 01 · Private Client Portal
Private Rental Owner Portal - Côte d'Ivoire
Private rental operations portal for a short-term rental owner, designed in French and FCFA.
Private client portal
How It Works
Operational messages arrive through Telegram, are classified and validated in n8n, then stored in Supabase. The owner also uses a private responsive portal to review performance and create or correct bookings, payments, and expenses.
Why It Matters
Owners of short-term rentals often need to keep track of reservations, mobile-money payments, and day-to-day costs in different places. When the information is scattered, it becomes harder to see what has been paid, what is still due, and how each property is performing.
Result / Output
- Telegram messages become validated payment, expense, or booking records.
- The owner can manage bookings, deposits, balances, payments, and expenses directly on the portal.
- Monthly and yearly performance is available on mobile and desktop.
- A visual PDF report can be produced from the portal.
Operational Impact
Instead of relying on scattered messages and spreadsheets, the owner has one place to review bookings, payments, and expenses. This makes it easier to check performance, correct a record when needed, and follow unpaid balances.
Design Logic
Telegram is used for everyday operational messages, while Supabase keeps the data organised in one place. The private portal gives the owner a simple view of the business in French and local currency.
Featured Workflow 02 · Tested V1
Property Management Automation Tool
A tested service fee tracking and reminder workflow for a single property, using Google Sheets, Telegram, n8n, and Supabase.
Working systemHuman approval
How It Works
Managers update tenant and payment information in Google Sheets, then use Telegram commands to sync changes, create a new service fee month, check payment status, and prepare reminders for review. n8n validates each action, while Supabase stores the records and current-month views.
Why It Matters
Service fee follow-up is often split between spreadsheets, chats, and memory. This makes it easy to miss unpaid fees, duplicate reminders, or act on information that changed after a payment. The workflow brings those steps into one controlled process.
Result / Output
- Managers update records without direct database access.
- Telegram handles sync, monthly setup, status, reminder review, and activity history.
- Paid or partial payments block outdated reminders. Approved reminders are sent manually.
Operational Impact
Creates a current-month picture of paid, unpaid, and partial service fees. It also prevents duplicate monthly records and stops old unpaid reminders from being approved after a payment changes.
Design Logic
Google Sheets gives the manager a familiar place to update records. Telegram controls the workflow, n8n validates the actions, and Supabase remains the source of truth. The manager continues to manage external communication.
Design Principles
How I approach practical AI automation systems.
These principles guide how I design tools that are useful, safe to test, and realistic for business and field environments.
01
Simple Interfaces First
I prioritize channels people already use, such as Telegram, SMS, WhatsApp, and spreadsheets, before proposing complex apps.
02
Structure Everyday Messages
Many operational problems begin when informal messages are lost through interactions. The goal is to turn those messages into tasks, records, summaries, and follow-up actions.
03
Keep Human Approval Where Risk Exists
For external messages, client communication, payments, or alerts, the system should prepare and log actions before sending them.
04
Build Logs Before Scaling Automation
Logs, status fields, and history tables make workflows easier to audit, debug, improve, and eventually hand over to clients.
05
Design For Real Operating Conditions
The best tool is not always the most advanced and complex interface. I design around connectivity, user habits, language, and the level of confidence in technology a team has.