ChatGPT Prompts for Power BI (DAX, Dashboards, Modeling)
20 copy-paste ChatGPT prompts for Power BI: DAX formulas explained, dashboard design, data modeling, M/Power Query, and the report-writing that turns dashboards into decisions.
In short: This page contains 16 copy-paste ready prompts, organized into 4 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.
DAX Formulas
4 promptsDAX Formula Generation
1/16✨ What it does
ChatGPT writes a DAX formula for [calculation] using [tables, columns, relationships], with an explanation of each function, common mistakes, and performance notes. Paste it into Power BI and check the result on a known row before you publish.
I need DAX formula for [calculation]. Context: [tables, columns, relationships]. Output: DAX expression, explanation of each function, common mistakes, performance considerations (CALCULATE filters, iterators). Verify in Power BI before deploy.
Pro tip: DAX is its own language. AI helps draft + explain; test in Power BI before deploy. Hallucinated DAX possible; verification mandatory.
DAX Optimization
2/16✨ What it does
ChatGPT rewrites [Paste slow DAX] for speed, explaining why it was slow, what should improve, and other approaches. Swap the new formula in and time the visual before you ship the report.
[Paste slow DAX]. Optimize for performance. Output: rewritten formula, why slow (iterators, filter context, table scans), expected improvement, alternative approaches. Performance matters at scale.
Pro tip: Slow DAX = unusable dashboard. Common culprits: nested CALCULATE, row-level iteration, unrelated filter context. Targeted optimization = 10-100x speedup.
Time Intelligence DAX
3/16✨ What it does
ChatGPT writes time-intelligence DAX for [YTD / MTD / vs prior period / rolling 12] using SAMEPERIODLASTYEAR, DATEADD, or DATESYTD, and flags date-table pitfalls. Confirm you have a marked date table, then test the measure against a known period.
DAX for time intelligence: [YTD / MTD / vs prior period / rolling 12]. Output: formula using DAX time-intelligence functions (SAMEPERIODLASTYEAR / DATEADD / DATESYTD), date table requirement, common pitfalls. Date table essential.
Pro tip: Time intelligence requires marked date table. Without it, time functions return wrong results. Always check: model has date table + marked + relationships.
DAX Debugging
4/16✨ What it does
ChatGPT debugs [Paste DAX returning wrong result] by walking filter context step by step and suggesting test variants. Run the tests in Power BI and keep the variant that matches the number you expected.
[Paste DAX returning wrong result]. Debug: what I expected, what it returns, where mismatch likely. Walk through filter context step-by-step. Suggest test variants. DAX debugging is filter-context puzzle.
Pro tip: DAX returning unexpected = filter context wrong. Walk filter context step-by-step. CALCULATE modifiers, REMOVEFILTERS, ALL — each changes context. Mental model first.
Prompts get you started. Tutorials level you up.
A growing library of 300+ hands-on AI tutorials. New tutorials added every week.
Data Modeling
4 promptsStar Schema Design
5/16✨ What it does
ChatGPT designs a star schema for [domain], with fact and dimension tables, relationships, grain, and slowly changing dimensions. Sketch the tables in the model view and fix grain before you load more data.
Design star schema for [domain — sales / HR / finance]. Output: fact table, dimension tables, relationships, granularity per fact, slowly-changing dimensions handling. Star schema > flat tables for performance + clarity.
Pro tip: Power BI optimized for star schema. Flat tables = poor performance + confusing measures. Star schema upfront = sustainable model.
Relationship Cardinality
6/16✨ What it does
ChatGPT recommends relationship cardinality between [tables], usually many-to-one, with why, cross-filter direction, and performance notes. Set that cardinality in the model and refresh to confirm your totals still match.
Help me decide relationship cardinality between [tables]. Many-to-one (recommended), one-to-one (rare), many-to-many (use carefully). Output: recommended cardinality, why, cross-filter direction, performance implications.
Pro tip: Many-to-many relationships = often modeling smell. Either dimension table missing or shared dimension needed. Avoid m:m where possible; design dimension out.
Power Query / M Language
7/16✨ What it does
ChatGPT writes Power Query M for [transformation needed], with a step-by-step explanation, query folding notes, and error handling, written as if M is new to you. Paste the steps into Power Query and check that folding still holds.
Power Query M for [transformation needed]. Output: M code, step-by-step explanation, performance considerations (query folding), error handling. M is unfamiliar to many; explain like first time.
Pro tip: M language poorly documented. AI helps draft + explain. Test in Power Query editor; verify query folds for performance (server pushdown).
Data Refresh Strategy
8/16✨ What it does
ChatGPT plans a refresh strategy for [dataset], covering schedule, incremental refresh, gateway needs, failure alerts, and capacity. Set the schedule and a failure alert so a stale dashboard does not surprise you.
Refresh strategy for [dataset]. Output: scheduled refresh frequency, incremental refresh setup, gateway requirements, monitoring + alerts on failure, capacity planning. Refresh failures = stale dashboards = lost trust.
Pro tip: Daily refresh fails silently for weeks = stale dashboards. Alerting on failure + monitoring SLA = users trust dashboards. Without monitoring = invisible decay.
Dashboard + Report Design
4 promptsDashboard Layout
9/16✨ What it does
ChatGPT lays out a dashboard for [dashboard purpose] aimed at [executive / operational / analytical], with 5 to 7 top KPIs, a filter pane, chart order, drill-down, and mobile notes. Build the page in that order and hide anything your audience will not use to decide.
Layout for [dashboard purpose]. Audience: [executive / operational / analytical]. Output: top-of-page key metrics (5-7 KPIs), filter pane, chart hierarchy by importance, drill-down navigation, mobile considerations. Design for the audience's decision.
Pro tip: Executive dashboard = 5-7 KPIs scannable in 30 sec. Operational = real-time + alerting. Analytical = drill-deep capability. Audience drives design.
Visual Chart Selection
10/16✨ What it does
ChatGPT picks the best Power BI visual for [data type + question], matching line to trend, bar to a short comparison, and table to a long list. Swap the current visual to that type and see if your question is easier to answer.
For [data type + question], best Power BI visual? Options: bar, line, scatter, treemap, KPI card, gauge, table, custom visual. Match visual to question, not data structure. "Show trend" = line. "Compare 5 things" = bar. "Show 100 things" = table.
Pro tip: Pie charts for >3 categories = unreadable. Gauges for KPIs not pacing toward target = confusing. Bar > pie for compare. Match visual to question.
KPI Card Design
11/16✨ What it does
ChatGPT designs a KPI card for [metric] with current value, versus target or prior period, a trend mark, color thresholds, and drill-through. Place the card at the top of the page and click through to confirm your detail view.
KPI card for [metric]. Output: current value, comparison (vs target / vs prior period), trend indicator (sparkline or arrow), color logic (green/yellow/red thresholds), drill-through. KPI cards prime real estate; design intentionally.
Pro tip: KPI without comparison = number floating. KPI with target + trend = decision-ready. Comparison context is what makes a number a KPI.
Color + Accessibility
12/16✨ What it does
ChatGPT audits Power BI colors for brand fit, color-blind safe palettes, avoiding red and green as the only signal, label visibility, and contrast. Change the theme and check the report in grayscale so your colors still work.
Power BI report color audit. Check: brand alignment, accessibility (color-blind friendly palettes), red/green for non-critical (avoid for color-blind), label visibility, contrast ratios. Reports used by everyone need accessibility.
Pro tip: Red/green encoding = invisible to 8% of men (color-blind). Blue/orange or shape encoding = accessible. Most BI defaults fail accessibility; conscious choice required.
Like these prompts? There are full tutorials behind them.
Learn the workflows, not just the prompts. 300+ easy-to-follow tutorials inside AI Academy — and growing every week.
Reports + Insights
4 promptsInsight Narrative from Dashboard
13/16✨ What it does
ChatGPT writes an insight narrative from [Paste dashboard data summary] for [audience], with three findings, what changed and a why hypothesis, actions, and what to watch next. Edit the hypothesis against what you know, then send the narrative with the dashboard.
[Paste dashboard data summary]. Write insight narrative for [audience]: 3 key findings, what changed + why hypothesis, recommended actions, what to watch next period. Dashboards show; narratives explain.
Pro tip: Dashboards alone = "look at this." Narratives + dashboards = "do this because of this." Decision-driven beats data-driven.
Variance Explanation
14/16✨ What it does
ChatGPT helps explain why [metric] changed [X%] versus [period], with candidate causes by dimension, data to verify, a stakeholder narrative, and what to do next. Check the split in the model, then share only the causes you can support.
Top variance: [metric] changed [X%] vs [period]. Help me explain: candidate causes (decompose by dimension), data to verify, narrative for stakeholders, action implications. Variance > 5% = explain proactively.
Pro tip: Stakeholders see variance > 5% in meeting = "why?" question. Explained proactively = analyst credibility. Surprised by question = looks unprepared.
Executive Summary Slide
15/16✨ What it does
ChatGPT turns [dashboard] into a one-slide executive summary with 3 to 5 KPIs, the key trend, the biggest concern, the biggest opportunity, and a recommended action. Paste that into a slide and leave Power BI for the people who will click with you.
Convert [dashboard] to 1-slide executive summary. Output: top KPIs (3-5), key trend, biggest concern, biggest opportunity, recommended action. Executive can't click into Power BI; static slide for them.
Pro tip: Power BI dashboards interactive. Executives often want static slide. Distill dashboard to slide = reach audience that won't click.
Anomaly Detection Setup
16/16✨ What it does
ChatGPT sets up Power BI anomaly detection for [metric], covering how to enable it on a visual, sensitivity, what you will see, and whether to alert or investigate. Turn it on one chart and review the first flagged points before you trust alerts.
Anomaly detection in Power BI for [metric]. Output: enable in visual, sensitivity setting, expected output, what to do with detected anomalies (alert, investigate). Power BI has built-in; underused.
Pro tip: Power BI built-in anomaly detection (AI-powered) = visible spikes/dips automatically. Most users don't enable; data tells them what changed.
Free tool
Prompt Optimizer
Turn a rough idea into a structured, professional AI prompt.
Frequently Asked Questions
Prompts are the starting line. Tutorials are the finish.
A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.
7-day free trial. Cancel anytime.
Related guides