The operator

A dashboard shows you numbers and leaves the thinking to you. The operator does the
reading, ranks what matters, and tells you what it would do — then does it once
you say yes.

What it does today, honestly

	Status
Watches every connected platform	✅ Available now
Works out what needs action, and what to leave alone	✅ Available now
Recommends the specific move, with the evidence	✅ Available now
Carries out the move	✅ On your approval
Runs unattended 24/7 and acts on its own	🔮 Roadmap — not shipped

You set what it may do. Nothing changes in a live ad account or store without you
approving it.

Recommendations — on Home

The operator surfaces what it found, ranked, each with the number behind it and a
way to go look:

Returns — Alexandria is returning 34% of its orders · EGP 41,200 · Check
Alexandria

That's the format on purpose: the decision first, the metric as evidence. It also
tells you when the right move is to do nothing — a quiet day is a real finding,
not an empty screen.

Ask it directly — Chat

Ask about the business in plain language, English or Egyptian Arabic:

"What should I do today?"
"Show my real profit"
"Fix my margins"
"Why is TikTok wasting spend?"

It answers from your numbers, not general advice — and it can pull the specific
order, campaign or shipment behind its answer. You can attach an image or a PDF when
that's easier than describing something.

AI Tools

Deeper, on-demand investigations rather than the daily briefing:

Tool	What it's for
Store Doctor	A full health pass over the store — what's broken, what's leaking
Demand Forecast	What's likely to sell, so you don't stock out or over-buy
Pricer	Pricing pressure-tested against your real margin after returns

These run deterministic math first and use AI only where judgment is genuinely
needed — so the numbers don't change between runs, only the interpretation. They
use your AI balance (in dollars), and the cost is shown before you run one.

How it decides

The same loop, every time:

Observe live data across ads, orders, deliveries and cash → Understand it as
one picture, profit after returns → Decide what deserves attention → Act on
your approval → Measure whether it worked → Learn from the outcome.

Two rules it follows:

Math before AI. The number comes from arithmetic; AI explains and prioritises,
it doesn't invent the figure.
No fake certainty. When confidence is low — thin data, poor COGS coverage — it
shows the number and softens the verdict rather than dressing a guess as a
decision.

Getting the most out of it

Fill in COGS. Every recommendation is ultimately a profit judgment; unpriced
products make the operator optimistic.
Connect the courier. Without returns data it can't tell a winner from a
campaign that only looks like one.
Give it a full week. Comparisons need a prior period to be worth anything.

Have it come to you instead of you opening the app:
Reports & alerts.