Topic hub
Semantic models, DAX, Direct Lake, refresh and performance, licensing, report design and migrations.
45 articles by the Solv Systems team
Power BI is the reporting layer most organisations already own, and the one most likely to be blamed when the numbers are slow, wrong or ignored. The fixes are rarely in the visuals. They are in the semantic model, the refresh design, the storage mode and the way workspaces and licences are set up.
This hub gathers our Power BI writing in one place. The fundamentals come first: what a semantic model is, how DAX and Power Query fit together, the difference between a dashboard and a report, and when a star schema matters. Then the operational guides: why refreshes fail, why reports are slow, how incremental refresh works, when Direct Lake beats Import or DirectQuery, and how to prepare a model for Copilot without breaking it for people. We also cover licensing in plain terms, the real cost of a Power BI project in the UK, the US and South Africa, migrations from Tableau, Qlik and Excel, and what each monthly Power BI update changes in practice.
Copilot in Power BI and the AI features that depend on a well-prepared model are listed here and in the Copilot and AI agents hub. Row-level security and workspace governance sit in the data governance hub, and Fabric capacity questions, including when Power BI Pro is enough, sit in the Fabric pricing hub.
Newest first.
A practical round-up of Power BI's October 2026 update: the Gantt visual, list slicer dropdown, surge protection change and file picker deadline, with status and who should care.
From October 2026, Power BI Desktop builds from March 2026 or earlier can no longer save and share to OneDrive and SharePoint. Who is affected and how to update.
Web modelling, DirectQuery on the web, TMDL view and a hosted Authoring MCP server mean you can now edit a Power BI semantic model in the browser. What works, and what still needs Desktop.
TEXTCONTAINS and TEXTSIMILARITY bring full-text search to DAX, APPROXIMATEDISTINCTCOUNT reaches Import and Direct Lake, and Direct Lake gets calculated columns. What each does and when to use it.
Power BI Q&A retires in February 2027. Q&A visuals will show an error, embedded Q&A has no Copilot replacement, and Copilot now runs from F2. What to inventory and what to move to.
PBIP is generally available and PBIR is the default. How we now build Power BI projects: Git from day one, agents in the build loop, and AI access decided model by model.
Power BI Projects (PBIP) are generally available and PBIR is now the default report format. What the formats are, how automatic conversion works, what to check first, and a migration checklist.
Power BI answers in Microsoft 365 Copilot are GA; Microsoft 365 Copilot inside Power BI is in preview. What each does, the new per-model AI-access setting, and a rollout checklist.
Microsoft is bringing agentic app creation to Power BI Desktop, and Fabric Apps to Power BI Pro and PPU at no additional cost. What was announced, its status, and where it fits against Power Apps.
Power BI Desktop is the free Windows tool where you build models and reports; the Power BI Service is where they are published, secured, refreshed and shared. How the two fit together.
Microsoft is retiring real-time streaming in Power BI - push, streaming and PubNub semantic models - with 31 October 2027 as the date. How to find what you have and move it to Fabric.
In Power BI, dashboards and reports are different objects with different jobs - and most people saying dashboard mean report. What each is for, how they differ, and which to build when.
Excel is not the enemy; it is the incumbent. Where Excel genuinely remains the right tool, where Power BI decisively wins, and the signals that say your organisation has crossed the line between them.
Report design is not decoration; it is whether users find, read and believe the numbers. The layout, visual-choice and cognitive-load practices we apply on every engagement.
Most KPI dashboards fail before design starts: wrong metrics, no targets, no owners. How to choose KPIs worth tracking, present them so they trigger action, and govern them so they stay true.
Business intelligence is how organisations turn data into decisions: the pipeline from source systems to reports people trust. What BI involves, and where AI changes it.
DAX is the language behind every Power BI measure: powerful, compact, famous for humbling Excel experts. Why filter context is the whole game, and how to learn it without pain.
Power Query is the shape-the-data layer inside Power BI, Excel and Fabric Dataflows: connect, clean, transform, load - repeatably. What it does, where M fits, and when to graduate to heavier tools.
Free, Pro, Premium Per User, Fabric capacity: Power BI licensing trips up almost every buyer. What each licence actually permits, the F64 rule that changes everything, and how to buy for your size.
Key influencers, decomposition tree, anomaly detection and smart narratives shipped long before Copilot and still carry real weight. What each does, where it shines, and the caveats that matter.
Copilot, Fabric IQ and data agents answer questions by reading your semantic model. A checklist for making it legible to machines: naming, descriptions, relationships, security and endorsement.
Dataset, model, report - three words often used interchangeably for different things. What the semantic layer actually is, why it exists, and what changes once you have a good one.
Realistic durations by scope, what actually drives the timeline, and the discovery answers that change an estimate most. The companion question to what it costs.
The skills that separate a report builder from a modeller, practical interview tasks that actually reveal them, and how to tell whether you need this hire at all.
Utilisation, realisation and matter profitability - the three numbers partners argue about, and why the practice management system alone cannot settle them.
There is no automated path - workbooks get rebuilt. What transfers, what does not, realistic effort per dashboard, and the licence arithmetic that usually drives the decision.
Which spreadsheets to move first, which to leave alone, and how to migrate without finance losing the control they actually rely on.
The one genuine architectural difference between Qlik and Power BI, why it catches migrations out, and how to handle it rather than pretend it is not there.
Eight causes account for almost every Power BI refresh failure. Here is how to tell which one you have, fix it, and stop it recurring next month.
Slow reports are usually diagnosed in the wrong order. Here is how to find out whether the problem is the model, the DAX, the visuals or the source - before changing anything.
Flat tables work until they do not. Why the star schema is the single highest-leverage decision in a Power BI estate, and how to build one that holds up.
Full refreshes stop working at a size most estates reach eventually. What incremental refresh actually does, when it earns its complexity, and the setup mistakes that quietly disable it.
A permanent BI hire is not the only option, and often not the right one. An honest comparison of retained analytics support against recruiting, including when hiring wins.
Licences don't create a data-driven culture - capability does. Why Power BI rollouts stall, and how role-based training, governed self-service and champions turn licence spend into daily use.
If skilled staff lose days every month copying, pasting and formatting the pack, you're paying analyst salaries for clerical work. How automated Power BI reporting gives that time back.
Finance, sales and operations each report a different figure for the same KPI - and every meeting starts with a debate. Why it happens, and how a governed Power BI semantic model ends it.
Power BI vs Tableau, side by side for 2026: modelling, visualisation, Copilot vs Tableau Agent, integration, governance, deployment, verified list prices, and which to choose by scenario.
Direct Lake, Import, and DirectQuery are Power BI's three data connection modes. Choosing the right one determines report performance, data freshness, and Fabric cost.
Microsoft Power BI is the world's leading business intelligence platform. Here is what it is, how it works with Microsoft Fabric, and how it can benefit your organisation.
Power BI projects in the US range from $3,000 to $300,000. Five real project profiles with USD cost ranges, licensing costs, and what drives the variation.
A frank assessment of the Microsoft Fabric and Power BI consultancies worth considering in the US in 2026 - who fits which project, and who doesn't.
A frank assessment of the Microsoft Fabric and Power BI consultancies worth considering in the UK in 2026 - who fits which project, and who doesn't.
Power BI projects in the UK range from £2,500 to £250,000. Five real project profiles with honest GBP cost ranges, licensing costs, and what drives the variation.
Power BI projects in South Africa range from R40k to R4m. Five real project profiles with honest cost ranges, licensing costs, and what drives the variation.
A frank assessment of the eight Microsoft Fabric and Power BI consultancies worth considering in South Africa in 2026 - who fits which project, and who doesn't.
Articles filed under another topic that also cover Power BI.
From a slow report to a full semantic model rebuild, our consultants fix the model, not just the visuals. Book a free 30-minute call and bring the report that worries you.