Business Brain
A shared layer of business context: your offers, customers, activity, goals, permissions, and the relationships between them.
Not a separate database. Not a replacement for your judgment.MIDAN AI / ACTIONS INSIDE YOUR BUSINESS COMING SOONقريبًاNot another chat.
Not another chat.
AI that can act.
Midan AI is being designed to turn business signals into reviewable actions inside your workspace: prepare a page, create a follow-up task, or update a record. After approval—or within a permission you set—it executes and logs the result. This experience is not available yet.
A direction, not a claim of fully shipped capability.
BUSINESS BRAIN / PREVIEWILLUSTRATIVE · READ-ONLY
One decision, five connected signalsContext window 05
MIDANOWNER
Build
Sell
Manage
Automate
Grow
PROPOSED ACTION / ILLUSTRATIVEThree pricing questions → prepare an offer-page update and create a follow-up task.Review → approve → execute in Midan → log the result
THE PROMISE
The owner stays in the loop. The system gets better at the work.
AI should make the connected system more useful without turning the business into a black box. Midan AI is being designed to make context visible, decisions reviewable, and action proportional to permission.
CLICK TO BUSINESSCOMING SOONقريبًا
Describe what you want to build. Review a connected first draft.
Talk through your offer, customers, and goals. Click to Business is planned to prepare a connected first draft across funnels, websites, forms, surveys, store, products, and automations. You review and adjust every part before launch.
How it works
01
Describe what you want to build
Start with the offer, audience, and outcome you have in mind.
02
Connect the business context
Give the first draft a place in your offers, customers, and goals.
03
Prepare a connected first draft
It can prepare funnels, websites, forms, surveys, a store, products, and automations.
04
Review and adjust every part
You edit what needs changing and decide what is ready before launch.
What you get
This is a planned experience and is not available today. AI prepares a first draft for owner review; it does not automatically complete or launch a business.- A connected first draft across funnels, websites, forms, surveys, store, products, and automations
- The owner reviews and adjusts every part before launch
- No automatic business completion or launch
- Planned experience — not available today
ARABIC VOICE NOTES COMING SOONقريبًاFrom a voice note
From a voice note
to a reviewable draft.
Midan AI will transcribe voice notes in Egyptian, Gulf, and Levantine Arabic and prepare a reply in the customer’s dialect for review.
VOICE NOTE EXAMPLE
م
0:14
PLANNED TRANSCRIPTION · FOR REVIEW
«مساء الخير، هو الكورس ده للمبتدئين؟ وبكام؟»
PLANNED DRAFT REPLIESIllustrative · nothing sent
01
02
03
THREE CONNECTED ROLES
Not a chatbot beside the work. Intelligence inside it.
Brand Voice
A working understanding of how your business sounds, who it serves, and what it should never promise.
You can inspect, edit, and reset the source of that voice.Contextual Copilot
Useful support inside the place where work happens: a next-step suggestion, a clearer summary, or a draft grounded in the record.
The recommendation stays attached to its source and its reason.VISIBLE LOOP · CLEAR PERMISSIONS
Explain before you execute.
Every meaningful action should arrive with enough context to make a real decision: what was noticed, what would change, who or what it affects, and what it may cost.
01
Notice
The system identifies a pattern across the work you have connected.
02
Explain
It shows the sources, reasoning, proposed change, and expected cost before execution.
03
Authorize
The owner or a named role chooses the permission level and boundary.
04
Record
The result, actor, inputs, and outcome stay in the activity log.
ACTION LEVELS
Permission is a product decision.
A suggestion can stay a thought, become a reviewed change, or run only within a boundary the owner defines.
01
SuggestSUGGEST
Shows a recommendation and the context behind it.
You decide whether anything happens.
02
ApproveAPPROVE
Prepares the exact change, recipient, or workflow for your review.
Nothing runs until an authorised person confirms.
03
Authorized AutoAUTHORIZED AUTO
Runs a narrow, pre-approved action inside a defined boundary.
You choose the scope, limits, and the way to pause it.
Owners set roles and accessible records; show any cost before a paid action; and can narrow, pause, or revoke access at any time.
IN THE REAL WORK
Small moments where context earns its keep.
The point is not to add another place to chat. It is to reduce the distance between a signal and a considered next move.
01
CUSTOMER FOLLOW-UPA lead asks the same question twice.
Midan AI can point to the conversation, the offer page, and the unanswered detail. It can draft a reply for review; it does not send one just because it noticed a pattern.
Grounded example · review first02
WEEKLY REVIEWA campaign gets attention but not bookings.
See the connected signals behind the suggestion: campaign visits, the booking page, and recent opportunity stages. The useful output is a clearer next question, not a confident guess.
Grounded example · review first03
TEAM HANDOFFA client task is waiting on an owner.
Summarise what has happened, show the record it came from, and propose an assignment that respects workspace roles. The owner can edit, approve, or leave it alone.
Grounded example · review firstWHAT IS TRUE TODAY
A clear line between shipped work and the direction ahead.
Midan AI is still being shaped. We will expand in measured stages, using real outcomes to decide when a capability is ready.
Connected business modules, permissions, activity, records, and owner-controlled workflows are the foundation AI is being shaped around.
Business Brain, grounded suggestions, reviewable drafts, narrow authorised actions, and transparent AI credits are product direction—not promises of availability today.
A MEASURED ROLLOUT
Trust is something we measure.
Midan AI should earn its place through useful outcomes, not a launch announcement. Each stage is observed before the next one expands the boundary.
Grounded suggestions
Measure relevance, source quality, owner edits, and time saved before anything acts.
Approval workflows
Measure approval confidence, reversals, permission clarity, and whether the explanation was enough.
Narrow authorized action
Expand only where outcomes, costs, safety, and owner control remain visible.
The activity log is part of the experience.Every suggestion, approval, authorization, cost estimate, and result should leave a legible trail.
Read the FAQs PRACTICAL QUESTIONS
Trust gets specific.
Is Midan AI available today?
The page describes the direction and the product principles. Midan AI is coming soon; the examples are read-only product thinking, not a claim that every capability is shipped.
Will it be able to change my business automatically?
Only where a workspace owner explicitly authorises a narrow action. The intended experience starts with suggestions and review, with permissions, limits, and a pause control visible.
What will it use as context?
Only the Midan work that the workspace makes available to it: records, activity, offers, conversations, and connected signals. The source and scope should be visible before a meaningful action.
How will I know why it suggested something?
Each useful suggestion is designed to show the relevant sources, the proposed change, affected records, and any estimated cost before you decide.