Power BI consulting across the UAE
Power BI and Microsoft Fabric consulting for finance, operations and leadership teams in Dubai, Abu Dhabi, Sharjah and the Northern Emirates. Delivered remotely, 7am to 7pm Gulf Standard Time, seven days a week, at a fixed price quoted in writing before any work starts.
Oakwood Group builds Power BI reporting for organisations across the UAE: the data model that joins your finance system, operational systems and spreadsheets into one reporting layer, the dashboards on top of it, and the KPI definitions that stop finance, operations and the board arguing about whose number is right. Where data volume or refresh frequency calls for it, the same work is built on Microsoft Fabric. Delivered in partnership with Roar Data, an Australian Power BI consultancy.
It is for organisations that have outgrown Excel month-end packs, that have a Power BI estate nobody trusts, or that are weighing up Fabric and want a straight answer on whether it is worth the capacity cost. Everything is delivered into your own Microsoft tenant, on licences you already hold or buy from Microsoft directly. We are a reporting practice, not a software reseller, and not tax agents, auditors or lawyers: where this page touches VAT, Corporate Tax or data-protection law it explains how reporting handles them, and nothing more.
What a Power BI consulting engagement includes
Five phases, each with one deliverable you can hold in your hand. The detail of each phase is on the service page; this is the shape.
Discovery and scope
A short call at no charge, then a look at your current reports and the systems behind them. Deliverable: a written scope naming sources, model, pages, users, security, training and handover, with the fixed quote against it.
Definitions
A working session with the four or five people who own the numbers, on Teams. Deliverable: a KPI dictionary — each measure defined once, with its source, its filters and who signed it off.
Data model
The star schema underneath every report: entities, dates, currencies, products, sites, customers, and the fact tables from each source. Deliverable: a documented semantic model in your workspace, refreshing on schedule, reconciled to the source system.
Reports
The dashboards and detail pages agreed in scope, with row-level security applied and tested. Deliverable: published reports in your tenant, reviewed with the people who will use them, and a sign-off against the KPI dictionary.
Handover
Training on your own reports and model rather than on a demo, plus the documentation to change them without us. Deliverable: a handover pack, and a managed-support option if you want refresh monitoring and small changes handled for you.
Abu Dhabi
Business intelligence consulting for Abu Dhabi runs the same way as anywhere else in the country. The reporting conventions do not.
Reporting conventions in government-linked organisations
A large share of Abu Dhabi organisations are government-linked or semi-government, or supply organisations that are. Their reporting has a recognisable shape: a board pack on a fixed monthly or quarterly cycle, a KPI cascade in which the organisation's strategic objectives break down into department and then individual measures, and an executive layer that is often bilingual. Power BI handles all three, with some design choices worth making early.
The board pack is the easiest to get wrong, because it is a document rather than a dashboard. Building it as a paginated report, or as a fixed set of pages exported to PDF on schedule, keeps the layout stable from one meeting to the next while the numbers come from the same model as the live dashboards. The KPI cascade wants a measure table with a parent-child hierarchy, so a director's view rolls up the same measures a team lead sees, rather than two separate calculations that drift apart. For Arabic executive labels, the honest position is that Power BI's right-to-left support is limited: text direction can be set in some visuals, but the report canvas does not mirror. The workable approach is a bilingual field-name layer in the model — an English and an Arabic caption per measure and dimension — with the layout designed left to right and the Arabic labels placed deliberately.
ADGM entities, the UAE Central region and procurement
An entity registered in Abu Dhabi Global Market is subject to the ADGM Data Protection Regulations 2021 rather than the federal law, which affects where personal data in a report may be stored and processed and what has to be documented about it. Azure has a UAE Central region located in Abu Dhabi alongside UAE North in Dubai; which one your tenant and capacities use is a Microsoft configuration decision, and the data residency section below explains how to check it. None of this is legal advice — an ADGM entity's data protection officer or counsel decides what the regulations require; we build the reporting to fit the decision.
Procurement in Abu Dhabi often runs through supplier registration portals and, for larger buyers, in-country value scoring. Those are your processes and your buyer's processes. Oakwood Group does not hold an ICV certificate, is not a registered supplier with any Abu Dhabi entity and does not claim to be an approved vendor of anyone; if your procurement requires a registered supplier, that is a conversation to have before the discovery call, not after the quote.
Where the Abu Dhabi weighting falls
Three sectors carry an Abu Dhabi weighting in the reporting we build. Real estate and property management, where Tawtheeq tenancy registration, community and service-charge accounting and a portfolio-to-unit drill path shape the model. Energy services, where contract-level margin, mobilisation and HSE measures sit alongside the finance numbers. And logistics around KIZAD and the AD Ports cluster, where throughput, dwell time and customs clearance join the freight and warehouse data.
Sharjah and the Northern Emirates
Manufacturing, free-zone trading and multi-entity groups with a plant in one emirate and a sales office in another.
Sharjah, Ajman, Umm Al Quwain, Ras Al Khaimah and Fujairah hold much of the country's manufacturing and a large part of its free-zone trading. Sharjah Airport International Free Zone and Hamriyah Free Zone in Sharjah, and RAKEZ in Ras Al Khaimah, are home to thousands of industrial and trading licences, and the businesses behind them tend to share one structure: a production or warehousing entity in the Northern Emirates, a sales or trading entity in Dubai, and a group that reports on both.
The reporting problems follow from that structure. Production data lives in the plant system — often an ERP module, sometimes a shop-floor spreadsheet — while sales orders and invoices live in the Dubai entity's books, and the group consolidates the two in Excel every month. A Power BI model with a proper entity dimension holds both sides, reconciles intercompany sales between them, and gives the plant a production dashboard and the sales office a margin dashboard from the same numbers. Manufacturing measures that recur in this work include yield and scrap, downtime by cause, production versus plan, cost per unit with landed material cost, and inventory ageing across the plant and the sales warehouse.
Because delivery is remote, an organisation in Ras Al Khaimah or Fujairah gets the same engagement as one in Dubai Internet City. Discovery, definitions and reviews run on Teams; the model is built against your systems through a gateway or a cloud connection; reports are published to your tenant. There is no travel component in the quote and no difference in cadence between emirates.
Free zone and mainland in one report
The standard UAE group structure, and what a data model needs so the group, each entity and each tax registration can all be read from the same tables.
Most UAE groups of any size are a mix of mainland and free-zone entities: a mainland LLC that can trade onshore, one or more free-zone companies that hold licences, staff or assets, and sometimes a holding company or a branch. Finance keeps separate books for each because the law requires it, then spends month-end consolidating them by hand. The reporting model has to respect the separation and still produce the group view.
- Entity dimension
- Every fact table carries the legal entity, and the entity dimension carries its jurisdiction (mainland, free zone, designated zone, DIFC or ADGM), its VAT registration and its Corporate Tax status. Group, entity and jurisdiction views are then filters on one model, not three models.
- VAT registrations and tax groups
- Entities may hold their own Tax Registration Numbers or be grouped under one. The model tags each transaction with the registration it reports under and with its supply type — standard-rated, zero-rated, exempt, out of scope, reverse-charge import — so the VAT summary in a dashboard reconciles to the return filed with the Federal Tax Authority. Reconciling is what the report does; filing and the treatment of any transaction remain your tax agent's work.
- Corporate Tax segregation
- A qualifying free-zone person may have qualifying income taxed at 0% and other income at 9%, which is an entity-level distinction that the finance team and its advisers make. The reporting job is to keep entity-level income segregated in the model so those figures can be read without re-cutting the data. The model does not decide what qualifies; that is explanatory only and your adviser's call.
- Intercompany eliminations
- Sales from the free-zone plant to the mainland trading company are real revenue at entity level and nothing at group level. The model flags intercompany counterparties so the group view eliminates them automatically, and a reconciliation page shows the two sides of each intercompany balance so a mismatch is visible before the auditors find it.
- AED base with foreign purchases
- The dirham is the reporting currency, but purchases arrive in USD, EUR, CNY and INR and GCC sales in SAR, QAR or OMR. The model stores transaction currency and amount, an AED amount at the rate booked, and a rate table so margin can be shown at booked, month-end or budget rates. Because the AED is pegged to the dollar, USD exposure is small; the exposure that moves margin is usually EUR, CNY and INR, and that is the one the dashboard makes visible.
Most of this finance model — the chart of accounts mapping, the entity and currency handling, the reconciliation to the ledger — is the same work as moving a group out of Excel consolidation. The migration page sets it out in detail.
Microsoft Fabric in the UAE — when it is worth it
Fabric adds a data platform underneath Power BI. Whether that is worth paying for by the capacity depends on what is failing today, not on the product roadmap.
Power BI Pro and Premium Per User are per-user licences: each person who builds or views pays a monthly fee, the model runs inside Microsoft's shared service, and data preparation happens in Power Query on refresh. Fabric changes the underlying platform. You buy capacity — an F-SKU sized in capacity units, from F2 up to F2048 — and inside it you get OneLake, a single data lake for the tenant; lakehouses and warehouses that hold data in open Delta format; pipelines and notebooks for moving and transforming data; Direct Lake, which lets a semantic model read those tables without importing them; and Power BI itself. On capacities of F64 and above, people who only view reports do not need a Pro licence.
The case for Fabric is a data engineering case: several sources that need landing, cleaning and joining before they reach Power BI, models that have outgrown per-user limits, refreshes that no longer finish, or a wish to hold a governed copy of the data that other tools can read too. The case against is cost and complexity where none of that applies. The table is how we make the call with a client.
| Stay on Pro or PPU if… | Move to an F-SKU if… |
|---|---|
| Models fit comfortably within the Pro (1 GB) or PPU (100 GB) model limits | Models exceed PPU limits, or refresh is timing out at the size you need |
| Scheduled refresh keeps up: 8 refreshes a day on Pro, 48 on PPU, is enough | Data has to be near-real-time, or pipelines need to run on their own schedule and feed reports as they land |
| Data preparation is manageable inside Power Query and one or two dataflows | Several sources need a staged bronze, silver, gold preparation layer with lineage and reuse across models |
| Everyone who views reports is already licensed, or the viewer count is small | A large viewer population makes F64 or above cheaper than per-user licences, or reports must be embedded for external users |
| The pain is report design, definitions or a poor model — Fabric will not fix those | You already hold a Premium P-SKU that is being retired and must move somewhere |
| Nobody in the organisation will own a data platform after handover | There is, or will be, someone accountable for the lakehouse and its pipelines |
How capacity cost behaves
Capacity is billed for the hours it runs, not for the number of people using it. A pay-as-you-go F-SKU can be paused when nothing needs to run and scaled up or down as the workload changes; a reserved capacity commits to a year at a lower hourly rate and cannot be paused. Workloads can burst above the provisioned capacity for short periods, with the excess smoothed over the following hours; sustained overuse leads to throttling, which shows up as slow or rejected queries at the moment people are opening reports. That is why heavy pipelines are scheduled away from the hours reports are read, and why the capacity metrics app is set up on day one rather than after the first surprising invoice. Premium per-capacity P-SKUs are being retired in favour of Fabric capacities, so an organisation on Premium today is choosing an F-SKU at renewal whether or not it adopts the rest of Fabric.
What the platform looks like
Sources Fabric Consumption
------- ------ -----------
ERP (BC, SAP B1, Odoo) → Pipelines / Dataflows Gen2
Payroll / WPS files → ↓
CRM, PMS, WMS, TMS → Lakehouse in OneLake (Delta tables)
Excel budgets, rates → bronze → silver → gold
Customs / bank feeds → (raw) (cleaned) (modelled)
↓
Semantic model (Direct Lake or import)
↓
Power BI reports, paginated board pack, ExcelBronze holds each source as it arrived, so a number can always be traced back. Silver holds it cleaned and typed, with the entity, currency and date keys applied. Gold is the set of tables shaped for reporting — the star schema — and it is what the semantic model reads. The same gold tables can feed more than one model, which is how a finance model and an operations model stop disagreeing.
Migrating an existing Power BI estate
An existing estate does not have to be rebuilt to move to Fabric. Assigning a workspace to a Fabric capacity keeps every report, semantic model, measure and security role working as it did. From there the migration is incremental: the data preparation currently inside Power Query moves into pipelines and the lakehouse, gold tables replace the queries one source at a time, and each semantic model is repointed at OneLake — keeping import mode where it works, moving to Direct Lake where the model is large or the data has to be fresh. Reports keep their existing models until the replacement reconciles to the same totals. If the estate is slow because the model is wrong rather than because the platform is too small, migrating it moves the problem rather than fixing it; that judgement is made first, and our article on when Power BI should be rebuilt rather than patched sets out how.
Residency in UAE regions
A Fabric capacity is created in an Azure region you choose, and OneLake data for that capacity is stored there. Microsoft publishes which regions offer Fabric and which workloads are available in each; UAE North is the region to check first for a UAE tenant, and availability should be confirmed at the time of purchase rather than assumed, because the list changes. Licensing for Pro, PPU and the F-SKUs, and how they combine, is covered on the Dubai hub page.
Data residency and protection
Three data-protection regimes apply in the UAE depending on where an entity is registered. This is what each one means for a reporting build, in plain terms and without advice.
Federal PDPL. The Personal Data Protection Law, Federal Decree-Law 45 of 2021, applies to entities in mainland UAE and in the ordinary free zones. For a reporting build the practical questions are which personal data — employee records, customer contact details, tenant identities — actually needs to be in the model at all, whether it can be reduced to an identifier, and who can see it once it is there. Row-level security and workspace roles are the controls; deciding what the law requires is your data protection lead's call.
DIFC. Dubai International Financial Centre has its own Data Protection Law 2020 and its own Commissioner. A DIFC entity is under that regime rather than the federal one, which is mainly a documentation and transfer question: where the data is processed and on what basis it leaves the jurisdiction. A tenant in Azure UAE North is inside the country; whether it is inside the DIFC's rules for a given dataset is something the entity's compliance function decides.
ADGM. Abu Dhabi Global Market has the ADGM Data Protection Regulations 2021 and an Office of Data Protection. As with DIFC, an ADGM entity answers to its own regime. The build accommodates whatever that regime requires of the data in the model — minimisation, access control, a record of what is held — but the requirement itself comes from your side.
Where the data lives, and what we do not do
Power BI stores model data in the home region of your Microsoft tenant, shown in the admin portal under the tenant's settings; for a UAE tenant that is normally UAE North. Fabric capacities are created in a region you choose. On-premises systems — an ERP on a server in the office, a SQL database behind a firewall — are reached through the on-premises data gateway, which you install inside your network and which makes only outbound connections; credentials are stored encrypted and the data passes through the gateway to the service without being held in between.
Everything we build is in your tenant, your workspaces and your capacity. We do not copy data out of your tenant into systems of our own, and access is through accounts you create and can remove. Working files during a build stay in a workspace you own.
Cadence, tools, access and handover
A Power BI engagement has one part that benefits from a room — the definitions session — and it runs well on Teams with the right four people in it. Everything else is built against your systems and delivered into your tenant, so the work runs the same way from any emirate. The delivery team works to Gulf Standard Time, 7am to 7pm, seven days; there is no daylight saving in the UAE, so the hours do not move through the year. During Ramadan, check-ins shift to reduced working hours, and go-lives are not scheduled in the last ten days of Ramadan or in Eid weeks unless you ask for them.
What we don't claim
You will not find client logos, star ratings or a partner badge on this site. Oakwood Group is a new practice in the UAE and we would rather show you how we work than borrow proof. What we can show you: our process, our pricing method, demo dashboards built on synthetic data that you can open right now, and the partnership under which the work is delivered.
Common questions about Power BI consulting in the UAE
Abu Dhabi, Sharjah, the wider Gulf, cost, residency and Fabric.
Do you serve Abu Dhabi and Sharjah?
Do you work in the wider GCC?
How much does Power BI consulting cost in the UAE?
Is my data kept in the UAE?
What is Microsoft Fabric and should we move to it?
What is Direct Lake?
Can you migrate our existing datasets to Fabric?
How do you keep Fabric costs under control?
How fixed pricing works
One price is fixed and published; the build is quoted. Either way the number is built from a written scope, and the quote is the price.
A short discovery call, at no charge
About 30 minutes, and a look at your current reports and the systems they draw from.
A written scope
Sources, model, report pages and visuals, users and security, training and handover — each named, so both sides know what is in and what is out.
A fixed quote against that scope
Quoted in writing before any work starts. The quote is the price.
Scope risk sits with us
If we find something worse than expected inside the scope, that is our cost. If you want something outside it, we quote that separately before doing it.
Payment stages
Set out in the quote alongside the scope, so the schedule is agreed before work begins.
What the numbers are
Reporting diagnostic — AED 4,950, fixed. A review of the reports you have now and the systems behind them, ending in a written action plan: what is slowing the reporting down, where the definitions disagree, and what to fix first. It is the same figure for everyone because the work is the same shape. That is the whole amount. Nothing is added at checkout. It is not a condition of a build — where a discovery call is enough to write a scope, we say so and skip it.
Build — from AED 15,500, quoted. Most builds land between AED 15,500 to AED 62,000: a single-source departmental build sits near the bottom, a multi-entity finance model in the middle, and a group joining an ERP to its operational systems at the top. A build is never bought from a price list, so treat that range as the start of a conversation about magnitude. The figure for yours is written against your scope before any work begins, and that written figure is the price.
We publish the method as well as the figures, because the method is what tells you where in that range your build sits. Licences are separate and are paid to Microsoft at their published rates.
Start with four quick questions
Tell us which systems your reporting has to pull from, which emirates your entities sit in, and what is slow or hard to trust today. A 30-minute call, no charge, follows — then a written scope and a fixed quote.