Data Platform Services

Microsoft Fabric & Data Warehouse Consulting in Singapore

The Governed Data Layer Behind Your Power BI, Built by a Finance-First Team

ITLink designs and builds Microsoft Fabric data platforms and data warehouses for Singapore finance and operations teams — consolidating ERP, CRM and spreadsheet data into one governed model that Power BI, TM1 and OneStream can all report from. Singapore-based, delivering since 1995, with Microsoft-certified Fabric Analytics Engineers (DP-600) and Power BI Data Analysts (PL-300) on the team.

Trusted by Singapore finance teams since 1995

Your systems
ERPCRMXero / NetSuite / SAPExcelHR & payroll
Microsoft Fabric
Pipelines & DataflowsOneLake / LakehouseData WarehouseSemantic model
Who reports from it
Power BITM1 / Planning AnalyticsOneStreamExcel
problems you face

Reporting Is Slow Because There Is No Warehouse Underneath It

Numbers That Don't Agree Between Systems

Revenue in the CRM, revenue in the ERP and revenue in the board pack are three different figures. Each report applies its own logic to its own extract, so reconciling them is a monthly job that never finishes.

Power BI Doing the Warehouse's Job

Hundreds of Power Query steps, refreshes that take hours and fail overnight, and a semantic model only one person understands. Power BI was asked to be the integration layer, and it isn't one.

ERP, CRM and Excel With No Single Source

The "master" customer list, product hierarchy and cost-centre map live in a spreadsheet someone maintains by hand. Every system has its own version, and finance is the one that has to stitch them back together.

Every New Report Starts From Raw Extracts

A new question from the board means a new export, a new cleanse and a new model. Nothing is reusable, history is lost when the source is overwritten, and last year's numbers can't be reproduced.

What We Deliver

Platform, warehouse and the reporting layer on top — designed together so the numbers agree by construction. Choose where to look.

Microsoft Fabric Platform

The Fabric tenant set up properly: sized for your workload, governed from day one, and deployable without hand-copying between workspaces.

Capacity sizing and F-SKU licensing advice
Workspace, domain and tenant design
OneLake and Lakehouse setup
Data Factory pipelines and Dataflows Gen2
Database mirroring (Azure SQL, Snowflake, Cosmos DB)
Purview governance and sensitivity labels
Deployment pipelines and Git integration
Capacity monitoring and cost control

Data Warehouse

A dimensional model built for finance questions — on Fabric Warehouse, Lakehouse or Azure SQL, whichever fits how you work.

Source-system audit and data profiling
Medallion (bronze / silver / gold) or classic staging design
Dimensional modelling and star schemas
Slowly changing dimensions and history tracking
Finance data marts: GL, AR/AP, sales, inventory, headcount
Period snapshots and close-locked balances
Reconciliation controls back to the ledger
Migration from Synapse, on-prem SQL Server or spreadsheet "warehouses"

Analytics Layer

The warehouse is only useful if people can report from it. We build the semantic model and connect it to the tools your teams already use.

Power BI semantic models on Direct Lake
Row-level security and certified datasets
Feeds into IBM Planning Analytics (TM1) and OneStream
Ready-made connectors for Xero, Business Central, QuickBooks, SAP, NetSuite and Yardi
Excel connected to the model, not to exports
Management and board reporting packs
Training and documentation for your own analysts

How an Engagement Runs

Five phases, from source audit to ongoing support. A first release — one source, one finance mart, one management pack — is typically live in 8–12 weeks; the design comes out of your systems and your close, not a reference architecture slide.

01

Discovery and Source Audit

Which systems hold which numbers, how clean they are, who owns the master data, and the reports that take longest today. We profile the actual data, not the data dictionary.

02

Architecture, Capacity and Cost

Lakehouse or Warehouse, which F-SKU, what runs on Fabric and what stays where it is. You get a written design and a monthly running-cost estimate before anything is built.

03

Build Pipelines and Warehouse

Ingestion from each source, the staging and dimensional layers, history tracking, and reconciliation checks that prove what lands in the warehouse matches the ledger.

04

Semantic Model and First Reports

The Power BI model on top of the gold layer, security roles, and the first management pack — the reports finance has been rebuilding by hand every month.

05

Handover and Support

Documentation, training for your analysts and administrators, then on-demand support at whatever level you want — from occasional help to a fully managed platform.

Why Work With Us

Most data-platform builds are led by engineers who have never closed a month. Ours are led by people who have.

1995
Delivering since
100+
Singapore businesses
4.8/5
Client rating

Finance-first modelling. Chart of accounts, fiscal calendars, multi-currency, intercompany, restatements — we model the warehouse the way finance thinks about the business, so reports don't need a workaround layer on top.

Singapore-based delivery. Your timezone, your regulatory context, your reporting deadlines. On-site when it matters, not a call scheduled around someone else's working day.

One team across the whole stack. We build the warehouse and the Power BI on top of it, and we implement IBM Planning Analytics and OneStream. The data layer is designed to feed all of them, not just dashboards.

Grant-aware scoping. Data and analytics projects can qualify for Singapore government support. We scope work so it fits the EDG grant where it applies — several of our recent data-platform engagements were delivered with EDG support.

Straight answers. If you don't need Fabric — or don't need a warehouse at all yet — we'd rather tell you during scoping than six months in.

Is Microsoft Fabric the Right Platform for You?

Not always, and we'd rather say so upfront.

Usually a strong fit

You already run Microsoft 365 or Azure and Power BI is in use
Three or more source systems feeding finance and operations reporting
Group reporting across entities, currencies or business units
Power BI refreshes that are slow, fragile or hitting model-size limits
An existing Synapse, Azure SQL or on-prem SQL Server warehouse due for renewal
?

Worth a conversation first

-
One accounting system and reporting that Excel handles fine — a connector package may be all you need
-
Already standardised on Snowflake or Databricks with no Microsoft estate — Fabric adds little
-
No budget for always-on capacity and no one to own the platform after go-live

Common migration paths we deliver

Azure Synapse / SQL DWFabric Warehouse
On-prem SQL ServerFabric Lakehouse
Power BI import modelsDirect Lake on OneLake
Excel / Access "warehouses"Governed finance marts

Microsoft Fabric & Data Warehousing in Singapore: Common Questions

What is Microsoft Fabric, and how is it different from Power BI Premium?
Microsoft Fabric is Microsoft's all-in-one analytics platform: data integration (Data Factory), a lakehouse and warehouse on OneLake, real-time analytics, data science and Power BI, all billed through one capacity. Power BI Premium was only the reporting layer; Fabric capacity includes Power BI Premium features plus the data engineering and warehousing workloads underneath it. If you have Power BI Premium per-capacity today, Fabric is its successor.
Do we need Fabric to have a data warehouse?
No. A well-designed star schema on Azure SQL Database or an on-prem SQL Server is still a good warehouse, and for a single-source, single-entity business it is often the cheaper option. Fabric earns its place when you have several sources, growing data volumes, Power BI models hitting refresh or size limits, or a wish to consolidate separate ETL, storage and BI tools onto one platform. We build both and will tell you which fits.
Lakehouse or Warehouse in Fabric — which should we choose?
Both store data in the same Delta format on OneLake, so the choice is about how your team works. A Fabric Warehouse is T-SQL first — full transactional SQL, familiar to anyone from a SQL Server background — and suits finance marts and governed reporting. A Lakehouse is Spark and notebook first and suits large or semi-structured data and engineering teams. Many of our designs use a Lakehouse for raw and refined layers and a Warehouse for the gold finance layer that Power BI reads.
How much does Microsoft Fabric cost in Singapore?
Fabric is billed as capacity, in F-SKUs from F2 upwards, priced per hour and available on pay-as-you-go or reserved terms. Capacity can be paused when not in use, which matters for month-end-heavy workloads. Reporting users on F64 and above don't need individual Power BI Pro licences. For a typical Singapore mid-market finance platform we would scope an F-SKU in the F4–F16 band — roughly S$700–S$2,900 a month on pay-as-you-go before reservation discounts; we produce a monthly running-cost estimate as part of the architecture phase before anything is built.
Can the warehouse feed TM1 or OneStream as well as Power BI?
Yes — that is usually the point. The gold layer is designed as the single source for actuals, so IBM Planning Analytics (TM1) loads its cubes from it, OneStream pulls trial balances and drivers from it, and Power BI reports from it directly. One reconciled set of numbers, three tools reading it.
Is our data stored in Singapore?
Fabric capacity is provisioned in an Azure region and OneLake data stays in that region. Microsoft offers a Southeast Asia region hosted in Singapore, and we provision there by default for Singapore clients. The region is recorded in the architecture sign-off, and for regulated clients we lock the tenant so data cannot be moved out of the region.
Is a Fabric or data warehouse project eligible for the EDG grant?
Data and analytics capability projects can fall under the Enterprise Development Grant's innovation and productivity categories, subject to eligibility and the project scope. We have supported EDG applications for analytics work, including data-warehouse builds for Singapore mid-market companies in logistics, property and manufacturing, and we scope the engagement so the application is straightforward. See our grants and funding page.
What happens after go-live?
Your team owns the platform; we stay available. Most clients keep us on an on-demand support subscription for pipeline monitoring, model changes and new reports, at whatever level suits — from a few hours a month to a fully managed platform.

Talk to a Singapore Fabric & Data Warehouse Team

A scoping call, not a demo. Tell us which systems hold your numbers and which reports take longest, and we'll tell you what a Fabric or data warehouse build would realistically involve — and whether you need one.