AI & Automation

Useful AI makes no noise. It removes the work nobody wants to do.

Artificial intelligence and automation for Algerian companies

Useful AI in a company is nothing spectacular. It sorts emails, prepares quotes, chases invoices, reads documents. We find the tasks in your processes that lend themselves to it, then automate them without exposing your data.

Try the demoDiscuss your project
  • Use cases chosen for their real gain
  • Human approval where it matters
  • Sensitive data that does not leave your control

Interactive demo

Write a message, watch it being handled

Type or paste a business email, or load a sample. Simple rules classify it, pull out the useful data and prepare a reply. A person approves before anything is sent.

Processing chain

Simple rules, run in your browser

  1. Classify
  2. Prioritise
  3. Extract
  4. Draft
  5. Approve

Quote request, priority: high. 7 fields. Reply awaiting approval.

Your message

205 / 1200

Or load a sample:

Processed in your browser. Nothing you type is sent.

Highlighted: the extracted data. Dotted underline: the words that decide the category and the priority.

Rule-based analysis

Category

Quote request

Words found: quote, budget

Priority

High

Words found: before 15 November

Extracted fields

7 fields

Action and reply

Proposed action

Create a draft quote in the ERP for 2,000 cartons, reference C40. Pass it to the sales contact concerned (city noted: Oran).

Prepared reply

Hello,

We have received your quote request for 2,000 cartons, reference C40. City noted: Oran. Deadline noted: before 15 November. Your quote is being prepared and will be sent to you by your sales contact.

Kind regards.

Awaiting approval. Nothing leaves without a person.

Simplified demonstration. The sample messages, references, amounts and contact details are fictional. It applies simple rules (keywords and text patterns) in your browser: nothing you type is sent anywhere, and no AI model is called on this page. A real project relies on a language model, which copes with wording these rules do not recognise.

What we see

What holds back AI adoption in companies

Nobody knows where to start

Everyone talks about AI, nobody says what it would actually do on Tuesday morning in your sales department.

Individual experiments, with no framework

Employees paste contracts and customer files into consumer tools. Nobody knows what has gone out, or where.

Repetitive tasks that wear teams down

Re-keying, copy and paste between two applications, identical answers to identical requests. Skilled time spent on work that is not.

Fear of a costly, vague project

Without a nearby example and without a figure, it is hard to justify an investment to management.

Our method

Our method for automating without losing focus

We start small, on one specific task, and we measure before extending.

  1. 1Identifying tasks that can be automatedWe review your processes with the teams and keep the tasks that are frequent, repetitive and well defined.
  2. 2Data classificationWe separate what can be processed by an online service, what must be anonymised and what requires a model deployed locally.
  3. 3Pilot on one use caseA first automation goes live on a limited scope, with human approval before each sensitive action.
  4. 4Measure, adjust, extendWe measure the time really saved and the errors, make corrections, then extend to other tasks if the result justifies it.

Deliverables

What you receive

  • A list of tasks that can be automated, ranked by gain and by difficulty
  • An AI usage rule for your teams: what is allowed, what is not
  • One or more automations in service, connected to your existing tools
  • Tracking: volumes processed, time saved, cases passed to a human
  • Documentation and training for the people who supervise

Who it is for

This service is for you if

  • Your teams spend hours on re-keying and repetitive replies
  • You receive more requests by email than you can handle quickly
  • You want to put a framework around the use your staff already make of AI tools
  • You have an ERP or a CRM and want to add automation to it without replacing it
Discuss your project

Why Gnosis

What you get by working with us

Advice independent of vendors

We are not tied to any software vendor or hardware manufacturer. The recommendation starts from your need, not from a catalogue.

A written scope before any commitment

Scope, priorities, budget and schedule are set down in writing before the first day of implementation.

Local delivery, by a team in Algiers

A team based in Algiers that knows the country's regulatory, human and infrastructure constraints.

Straight talk on what is feasible

If a project does not have the means to succeed, we say so before it starts.

Frequently asked questions

What we are asked before we start

Which tasks should we start with for AI?

Those that are frequent, repetitive and clearly defined: sorting and answering routine emails, preparing quotes, chasing invoices, reading purchase orders or documents. The gain there is quick to measure.

Are we allowed to send company data to ChatGPT or Gemini?

It depends on the data. Personal data falls under law 18-07, and some information is confidential by contract. We classify your data, anonymise what needs to be and deploy a local model when necessary.

Will AI replace our staff?

In the projects we run, it removes repetitive tasks and leaves teams with judgement, relationships and unusual cases. Human approval is kept on actions that commit the company.

Do we have to change our software to automate?

No, in most cases. Automation plugs into what you already use: email, ERP, CRM, spreadsheets. That is in fact the condition for it to be adopted.

Further reading

Our articles on the subject

Our other services

All services

Start with a task, not a strategy

Tell us the most repetitive task in your company. We will tell you whether it can be automated, and how.

Discuss your project