The Majlis of AI

Counsel for AI in regulated institutions

We advise, we train and we build with you.

We help banks, telcos, governments and other regulated institutions across the GCC put AI into production fast. Your board, your regulator and your teams agree the controls once. Then every model and agent ships at full speed.

Book a conversation See what we offer

A control, written as code

control:     GOV-007
obligation:  Human review of every adverse credit decision
owner:       Head of Retail Credit Risk
applies_to:  agent/onboarding-assistant
rule:        route_to_human if decline
evidence:    tests/gov_007.feature
Illustrative · owned and tested in your own Git
Fig. 01 · PlanTwo tiers · one centre Everyone agrees on the centreOpen side · the way out

What we offer

We advise. We train. We build.

One practice for the whole path: decide what to build, teach your people to run it, and build it with them in your own tenant.

  • #1 · Majlis Advisory

    We advise

    Strategy, operating model, architecture and governance-as-code. We help you make the AI decision fast, with one owner and a control your regulator can test.

    • Decision sprint
    • Programme design
    • Design authority
    Book a decision sprintSee the advisory

  • #2 · Majlis Academy

    We train

    Short, custom programmes, rebuilt as fast as AI moves, on a core of regulation, architecture and method that does not. Your people adopt in weeks and keep the knowledge current themselves.

    • ADLC bootcamp · 4 days · open now
    • 3–4 day formats, never longer
    • Rebuilt every cohort on a solid core
    See the ADLC bootcampAll courses

  • #3 · AI Factories

    We build

    We design, build and run AI factories with your engineers: software, governance, cloud, data, MLOps and training. In your tenant, in your country, owned by your team.

    • Your tenant, your country
    • Governance and MLOps built in
    • Owned by your engineers
    See AI FactoriesStart a build
Start anywhere. Most institutions start with one decision sprint.contact@majlisofai.com

How the Majlis works

Decided once. Shipped at full speed.

The slow part of regulated AI is not the model. It is the decision nobody owns. We name the owner, write the control as a test, and let the release gate do the rest.

majlis · retail-onboarding · decisionsIllustrative

D-03 · Owner: Head of Retail Credit Risk

Release the onboarding agent to retail customers

Approved
  1. W01Owner named
  2. W03Architecture drawn
  3. W06Controls as tests
  4. W10Release gate · pass
  5. W10+In production
W06  GOV-007 human review on declines ........ test written
W08  GOV-001 data lineage recorded ............ pass
W09  GOV-004 model card is current ............ pass
W10  owner sign-off · Head of Retail Credit Risk  signed
W10  release gate ............................. PASS → production

Majlis method · decision workspaceIllustrative. The weeks are examples, not a commitment.

Governance is the seatbelt, not the brake. Agree the controls once. Then every release moves at full speed.

We advise, we train and we build with youOne practice, three ways in

The Tao

Methods change. These nine principles stay. Our product is not software. Our product is capability: institutions that can design, run and govern Cloud and AI by themselves.

  1. 01Systems, not silos

    We design the whole system, then the part. Frameworks are lenses, not cages.

    We rejectPoint solutions sold as transformations.

  2. 02Regulated by design

    Regulation is a design input, not a final gate. Controls are proportional to risk and run as code.

    We rejectCompliance theatre, and innovation that cannot pass an audit.

  3. 03Intentional foundations, emergent products

    We plan the runway centrally. Products emerge from long-lived teams.

    We rejectPoC factories, and architecture as a bottleneck.

  4. 04Codify everything

    If it is not in version control, it does not exist. Decisions, policies, documents and courses are code.

    We rejectKnowledge that lives only in slides and inboxes.

  5. 05Clarity is a control

    Ambiguity is a risk. We write in Simplified Technical English and draw in ArchiMate.

    We rejectJargon that hides weak thinking.

  6. 06Map before you build

    A Wardley map first. Industry standards such as BIAN and TM Forum ODA next. Invent only what differentiates.

    We rejectReinventing a commodity.

  7. 07Capability, not dependency

    We succeed when you no longer need us. What we design with you, we teach your people to run.

    We rejectBlack-box delivery and lock-in.

  8. 08The circle

    We decide where everyone can see each other: executives, engineers, risk and compliance in one room.

    We rejectArchitecture handed down from above.

  9. 09Pragmatism

    Vision anchored in experience. Principles guide; they do not bind.

    We rejectDogma, including our own.

We use the Tao to accept work. If an engagement asks us to work against three or more principles, we reshape it or we decline it.

we advise, we train and we build with you
  • Novel
  • Series
  • 5 volumes

Books

Buy the novelThe series
Cover of The Majlis of AI: A Tech Novel of the New Geopolitical Council, by Roberto Barbosa

Published on Amazon

It started as a novel about the room where AI is decided.

The Majlis of AI is a tech novel of the new geopolitical council, where innovation, enterprise and politics meet. Every stakeholder has a seat and a voice. Laws, regulations and policies also have a seat.

From one book to a practice

  1. NovelThe story and the method
  2. DialoguesArticles and the letter
  3. AcademyBriefings to bootcamps
  4. AdvisoryArchitecture and operating model
  5. LabsAI services

Collaborations · with Márcio Parente

The Regulated AI Operating Model

Five books for the people who must run AI inside a regulated institution.

1The Adoption GapAmazon 2The Governance SpineAmazon 3The Platform FoundationAmazon 4The Agent Operating ModelControl planes, wranglers, MCP governance and model inventoryAmazon 5Building and ShippingSDLC, data governance, change control and the roadmapAmazon

Dialogues

Short pieces for architects and executives in banking, telco and government. Each one starts from a framework you already know.

All articles on LinkedIn

    Newsletter · Majlis Dialogues

    One issue a month. One decision in each.

    A short letter for people who must sign off on AI in a regulated institution. Read it in ten minutes. Bring it to your next steering committee.

    Subscribe on LinkedIn
    • One framework, drawnA BIAN, ODA or TOGAF view with the one change that matters marked.
    • One regulatory signalWhat changed for AI in GCC financial services, telco or government, in plain words.
    • One question for the boardA question to put on the agenda, with the answer we would expect to hear.

    Contact

    Bring us the AI decision that is slowing you down.

    We start with one decision and move it to production. Write to us about advisory work, the Academy, or a briefing for your board.

    Email

    contact@majlisofai.com