Skip to content

For SMEs

Automation and AI for small and mid-sized companies, with clear criteria.

No grandiose digital transformation and no yet-another-tool: proportionate systems that bring order to how your team works.

For businesses and teams that keep the day-to-day running on WhatsApp, calls, notes, spreadsheets or an ERP — and too much manual effort. Any starting point is fine.

Criteria

When automation is worth it. And when it isn't.

Half of our job is saying no. This is the test we apply before proposing anything.

Worth it when…

  • The process repeats every week and eats hours.
  • The rules can be written down: if this happens, do that.
  • The data it needs already exists, even if scattered.
  • There is a real bottleneck, not an occasional annoyance.
  • An error can be caught and fixed with human review.

Not worth it when…

  • The process changes every week: stabilise it first.
  • The volume is so low it takes minutes by hand.
  • Nobody can write the rule: the problem is definition.
  • The only reason is that “everyone is doing it”.
  • The process is broken: automating chaos only speeds up chaos.

Common mistakes

What we usually find when we arrive.

Buying the tool before defining the process.
Automating a process nobody has stabilised.
AI pilots with no owner, no data and no success criteria.
Systems that depend on a single person (or a single vendor).
Measuring cost in licences only, not in the team's hours.
Trying to do everything at once instead of starting where it hurts.

How we think about it

Useful AI, proportionate automation, honest return.

Useful AI

AI goes where there is repetitive text or data and a person who reviews: classifying, extracting, drafting, answering from your documentation. Not deciding on its own.

Proportionate automation

Control grows with impact. An internal classification doesn't need the same rigour as something that touches customers. No over-engineering, no fragile pieces.

Honest return

Return is measured in hours recovered, errors that stop happening and decisions made with data. We don't promise numbers before the diagnosis.

Typical projects

How it usually starts.

Four common starting points for an SME or a small business. None requires a big programme: each piece works on its own and prepares the next.

Paperwork that files itself

Invoices, orders, requests or quotes that arrive by email or WhatsApp and end up where they belong — an ERP or a clean sheet — with human validation where it matters.

Data-entry hours recovered and fewer copy errors.

An assistant over your documentation

The team asks in plain language and gets answers based on your procedures, price lists or catalogues — not the internet.

Fewer interruptions for the person who “knows everything”.

Reporting that builds itself

Sales, operations or cash figures consolidated every week without copy-paste, with alerts when something drifts out of range.

Decisions on fresh data, not last month's.

Customer replies with clear criteria

Frequent enquiries, appointments and reminders handled on WhatsApp or email, in your voice and with your data, reviewed before sending.

Shorter response times without losing control.

How we work

Small, useful, and it stays in-house.

The same D5 method we use on projects, at SME scale.

Diagnosis

We understand the business and pick where to start.

One piece first

One concrete process, actually working, in weeks.

Human review

The team controls what goes out; nothing decides alone.

Handover

Documented and explained so you don't depend on us.

FAQ

Questions SMEs ask us.

Are we too small for this?

If a process repeats and eats hours, no. Projects are sized to the real problem; many useful pieces are deliberately small.

Do we need our data in order before starting?

No. Ordering the essential minimum is part of the work. Starting with one concrete process is usually the best way to order the data that process needs.

How much does it cost?

It depends on scope, and we don't make that up: the diagnosis exists precisely to size it. You start with small pieces that carry their own value, not a big closed programme.

What if we already use AI on our own?

Even better: there's experience to build on. The usual step is ordering that use — criteria, limits, sensitive data — and turning what works into a system.

Does this replace people?

That's neither the goal nor the usual outcome. It removes repetitive work and keeps review and the final say in the team, which gets its time back for what actually needs people.

If the day-to-day runs on manual effort, let's start with a diagnosis.

A 30-minute call, no commitment, no sales pitch.