Digitalisation in business
Why this is not an IT subject, and why it can no longer wait

1. A worn word, a subject that is not

Digitalisation, digital transformation, dematerialisation : the vocabulary has served in so many presentations that it no longer triggers much. Many directors now hear it as a seminar word.

The question behind it has nonetheless stayed very concrete. Here it is, without the vocabulary : does your company know, at any moment, what it has sold, what it owes, what is owed to it, and where the documents proving all of that are ?

If the answer takes three days, four people and the opening of a binder, the subject is not behind you. It is ahead.

2. What digitalising does not mean

Three misunderstandings are expensive, because they cause spending without change.

It is not buying software.

Software laid over a vague process does not clarify the process : it reproduces the vagueness, faster and at greater cost. The tool never decides in place of the organisation.

It is not scanning.

A PDF that cannot be searched, cannot be linked to an accounting entry and cannot be produced as evidence is still paper — with a size in bytes on top. Digitising a document does not make it data.

It is not an IT project.

It is an organisational decision with IT consequences. The order matters : when you reverse it, the board ends up arbitrating configuration questions instead of business ones.

3. Four reasons to take it seriously

a. The obligations are going digital whether you follow or not

Online filings, electronic payments, exchanges with the authorities, traceability requirements on supporting documents : tax and social administrations across the OHADA area are moving towards dematerialised procedures. That movement asks nobody's opinion.

A company that keeps its accounts offline is not exempted for all that : it produces by hand, under pressure, what the system expects in structured form. The cost of the gap does not disappear, it simply moves to another budget line — and it is paid in hours.

b. Re-keying costs more than the software

A figure entered twice always ends up diverging. On the day it diverges, nobody knows which of the two versions governs, and an internal investigation opens over a discrepancy of a few thousand francs.

That cost appears on no invoice. It sits in the closings that slip, the variances to justify, the audits where the requested document cannot be found, and the year-end weekends. It is real, it recurs, and it almost always exceeds the price of the tool that would have prevented it.

c. The company's memory must not fit inside one head

Ask yourself honestly : how many processes in your organisation are genuinely known to one person only ? The day that person leaves, the knowledge leaves with them, and their successor spends six months reinventing what already existed.

A tooled process is a written process. That is often the first visible benefit of a digitalisation project, and rarely the one put forward to justify it.

d. AI only works on clean data

This is the most current reason. Every company wants artificial intelligence. None will get anything out of it from data scattered across a notebook, three spreadsheets and an inbox.

Digitalisation is not an alternative to AI : it is its precondition. The organisations getting results from these tools today are the ones that put their data in order first. Those that put nothing in order buy an engine and discover there is no fuel.

4. The three mistakes that sink a project

Starting with the tool.

Software is chosen before anyone has described what is meant to be achieved. The project then becomes a configuration exercise with no stopping criterion : nobody can say when it is finished, because nobody wrote down what success would look like.

Switching everything at once.

A complete replacement, over one weekend, with no way back. That method turns the smallest surprise into a board-level crisis. A narrow scope genuinely in production is worth more than a complete scope in permanent testing.

Forgetting the people.

A tool nobody has been taught to use is a tool nobody uses. Teams carry on with the old method in parallel « just to be safe », and the company pays for two systems to get one result. Training is not the last line of the quote, it is a condition of going live.

5. Where to begin

Five steps, in this order. This is the frame we follow on our projects, set out on the Our method page.

1. Look at what actually exists. Not the process described in the manual : the one the teams apply, workarounds included. It is the workarounds that tell you where the current system hurts.
2. Pick a scope that hurts. The most expensive process, not the easiest one. A painless first project convinces nobody and funds nothing afterwards.
3. Write down what « it works » means, before starting. A closing deadline, a variance rate, a number of re-keying hours. Observable figures, decided in advance and by the business.
4. Migrate the data. The step everyone underestimates and the one that decides the outcome. A botched migration gets a good tool rejected within three weeks.
5. Train, then switch off the old system. In that order, never the reverse. As long as the old system stays reachable « just in case », the migration has not happened.

6. The subject is not the technology

A well-organised company that digitalises becomes fast. A disorganised company that digitalises becomes rapidly disorganised. The tool amplifies what it finds ; it does not correct it.

That is why we always start with a diagnosis, never with a product demonstration. What we are looking for at that first stage is not the list of your software : it is the precise place where your organisation loses time, money or evidence.

If you want to see what the result looks like once the work is done, our three products — SANKOFA for the audit file, VISA-GED for tax certification and documents, HAKILI for accounting and management — are applications of that same frame to three specific trades.