Complete mapping
Your real workflows, observed end to end — where time is lost, quantified.
AI operations cabinet — MENA · Africa · Europe
AI must prove
AI must prove
its worth.
Before you buy another tool, get a clear verdict: where AI earns your budget, where it does not, and what to launch first.
Transparent method
Confidential by default
Deliverables owned by the client
Naciri Assistant
"Where AI earns your budget — and where it does not."
The diagnostic
A diagnostic that feels
A diagnostic that feels
like a real product.
Not a generic PowerPoint report. A workspace where every use case is mapped, scored and decided — on your real data.
Use-case scoring
6 cases observed · ranked by net value
Use caseVerdict
Client quote responses
Launch
Invoice follow-ups
Launch
WhatsApp customer support
Pilot
Stock forecasting
Not yet
Application screening
Not yet
Justified cases
Recoverable hours / monthest.
≈ 120 h/month at maturity, across the 3 selected cases (estimate)
A verdict for each use case
Every opportunity scored on observable criteria: launch, pilot, or decline.
A costed 90-day plan
The first workstream, its cost, its owner, and the proof threshold before going further.
Beyond the back office
Example — fictional data
AI-ops reaches sales, too.
Lead routing, pipeline scoring, sales follow-ups: the same method applies to sales operations — on a second, separate fictional client.
Groupe Ménara — example · sales casesVerdict
Inbound lead routingLaunch
Sales pipeline scoringLaunch
Assisted sales follow-upsPilot
Cold qualificationNot yet
Sales workflow automation — no media buying, no advertising.
The method
Three phases. A decision
Step 01
Step 02
Step 03
Three phases. A decision
dossier at the end.
No questionnaire, no promise. We observe, we score, we decide — and the path takes shape.
Observation
We observe your workflows, your exceptions and your existing systems — from documents and interviews, not from a questionnaire.
Scoring
Every use case is scored on observable sector criteria — a transparent method, no black-box score.
Verdict
A decision dossier: what to launch, what to decline, in what order — and the proof threshold before committing further.
The deliverable
The report you
The report you
receive.
An interactive dossier, structured in four sections. Browse an example below.
Example — fictional data
Workflow mapping
The real journey of an order, observed — and the two points where time is lost.
2 high-potential stepsQuoting and invoicing concentrate most of the administrative time.
Friction pointsManual re-entry, email back-and-forth, follow-up delays.
Stable stepsProduction and delivery do not justify AI intervention at this stage.
Use-case scoring
Six cases scored on observable criteria: volume, repeatability, data quality, risk.
Use caseScoreVerdict
Client quote responsesLaunch
Invoice follow-upsLaunch
WhatsApp customer supportPilot
Photo quality controlPilot
Stock forecastingNot yet
Application screeningNot yet
Verdict by use case
Three clear decisions, a priority order, and what we decline — with the reason.
Assisted quotes
LaunchHigh estimated net gain, low risk. First workstream of the plan.
Effort LowPriority 1
Invoice follow-ups
LaunchFast ROI, controlled scope, data already available.
Effort LowPriority 2
WhatsApp support
PilotTo be tested on a segment before scaling. Response quality to validate.
Effort MediumPriority 3
Stock forecasting
Not yetInsufficient historical data. To reconsider next quarter.
Effort HighPriority —
Launch plan · 90 days
Four milestones, an owner per workstream, and a proof threshold that decides what comes next.
Framing & data access
Workstream 1 — Assisted quotes
Workstream 2 — Follow-ups
Measurement & proof threshold
At the end of the 90 days: a measured result on the first workstream. If it does not reach the threshold, we stop — that is the rule.
Three deliverables
Three documents. Nothing
Three documents. Nothing
theoretical.
Documents built on your real data — not generic slides.
Mapping
Example · fictional data
Workflow mapping
Your real processes, observed — not declared. Where time is lost, quantified by your own data.
Scoring
Example · fictional data
Verdict by use case
Every opportunity scored on observable criteria: launch, pilot, or decline. With the justification.
90-day plan
Example · fictional data
90-day launch plan
The first workstream, its cost, its owner, and the proof threshold that decides what comes next.
Our offerings
One entry point.
One entry point.
Four possible paths.
P1
On requestOpen →
P2
On requestOpen →
P3
On requestOpen →
P4
On requestOpen →
P5
On requestOpen →
Diagnostic Entry point · on request
The starting point of every engagement: the documented verdict on your use cases.
Monthly engagement
Continuous steering of the workstreams validated by the diagnostic.
Managed agents
Automations operated for you, with a duty to prove their value.
Team training
Your teams made self-sufficient on the selected tools.
Fractional AI leadership
A part-time AI lead, on the client side.
Why a firm
Too many tools promise.
Too many tools promise.
Few are accountable.
Most AI budgets fund subscriptions, not results. We take the problem the other way round: proof on your real workflows first, the tool second — and sometimes the verdict is to buy nothing.
Commitments
01Paid for the diagnostic, not for the sale
02Confidential by default
03Declared to the CNDP
04You own everything
What we hold ourselves to.
We sell no tools.
We are not here to sell you a tool. The diagnostic stands even when its conclusion is to buy nothing.
Your data stays with you.
Your workflows and your data stay with you. Every engagement is covered by a confidentiality agreement.
Processing declared to the CNDP.
Our processing is declared to the CNDP under law 09-08. Personal data never leaves the declared scope.
You own everything.
Every deliverable and every automation belongs to you — portable and reusable.
Why trust us
Proof, not promise.
No badges, no promise of results. Four verifiable commitments — and a practitioner behind them.
Paid for the diagnostic, not for the sale
The firm's fee does not depend on the tool recommended. A low score stays a low score, even when it means selling nothing.
Transparent method
Visible scoring criteria, explicit weightings — no black-box score. You see exactly why each use case is scored the way it is.
You own everything
Every deliverable and every automation belongs to you — portable, reusable, with no hidden rent.
15 years of production systems
A practitioner who has operated automated systems in production for fifteen years.