System Automation in Rwanda

What is system automation, and who needs it

It’s identifying repetitive tasks, then building scripts, scheduled jobs, integrations or low-code flows that do them reliably without a person. It suits finance teams producing the same reports every month, operations teams copying data between systems, programme teams compiling monitoring data, and anyone sending the same messages over and over.

The problem it solves

Repetitive work is expensive in ways that don’t show up on a budget. Hours go into copying, formatting and checking. Errors creep in when people are tired or rushed. Tasks are delayed when the one person who does them is away. Automation removes that work, makes results consistent and frees time for tasks that need human judgement.

Good candidates for automation

• Reports assembled from the same sources every week or month.
• Data transfers between systems, such as orders into stock or payments into accounting.
• Reminders and alerts: overdue invoices, expiring contracts, low stock and missed deadlines.
• Document generation: letters, certificates, quotes and receipts from templates.
• File handling: renaming, sorting and filing incoming documents.
• Checks: finding duplicates, missing information or values outside agreed limits.

Designing automation people can trust

Start from the process. Map it with the people who do it, and simplify before automating.

Handle exceptions. Anything unusual stops and goes to a person, rather than being guessed.

Log everything, so it’s clear what ran, when and with what result.

Alert on failure to a named owner.

Keep it changeable. Rules stored where they can be updated without rewriting code.

Measure time saved and errors avoided, before and after the change.

Use what you own where it’s enough, such as Microsoft 365’s automation tools, and build custom only where needed.

Measuring the benefit

Before building, we record how long the task takes and how often errors occur. After a few weeks of automation, we measure again. Those figures show whether the automation was worth it, justify the next one, and catch cases where the time saved was smaller than expected because the process still needs manual checks.

Examples of what automation can do

A monthly donor report assembled from three spreadsheets in minutes instead of two days. Stock levels checked each morning, with a purchase request drafted when an item falls below its reorder level. Payment reminders sent automatically a week before invoices are due. Incoming scanned documents named, sorted into the right folder and logged.

What the work involves

1. Find candidates and estimate time and errors for each.
2. Prioritise by value and effort.
3. Map and simplify the chosen process.
4. Build the automation with logging, alerts and exception handling.
5. Test alongside the manual process for a few cycles.
6. Switch over, measure results and move to the next task.

What’s included and what’s not

Typical timelines and what drives cost

A single automation using existing tools can take days; custom automations spanning several systems take weeks. Cost is driven by the number of systems involved, data access, the complexity of rules and exceptions, volume, and monitoring requirements.

Equipment typically used

Microsoft 365 automation and forms tools where suitable, scheduled scripts and jobs, APIs and integrations as covered under API development and integration, document templates, and logging and alerting.

Common mistakes we see

Automating a broken process. No handling for unusual cases. Automations that fail silently. One person who knows how it works. Rules hard-coded so every change needs a developer. Big programmes that deliver nothing for months. And no measurement of whether time was actually saved.