Excel with Copilot is cheaper and simpler than Power BI for most small business dashboards: there is no capacity to buy, the data stays in a workbook, and Copilot drafts charts and PivotTables on request. Power BI is better when data must refresh unattended, many people need the same live view, or the data outgrows a workbook.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowCost is where people go wrong, because Power BI looks cheaper on its price list ($14 a user) while Copilot inside it needs paid capacity, and Excel looks free until you notice Copilot needs a Microsoft 365 licence that includes it. The comparison below counts both, then looks at effort, refresh and sharing, which decide most real cases.
How much do Power BI and Excel with Copilot cost a small business each month?
For an illustrative four-person firm, Excel with Copilot costs about $21 a month for the builder, against about $319 a month for Power BI Pro with Copilot once Fabric capacity is added.
Consider an illustrative four-person removals firm that wants a weekly dashboard: jobs completed, revenue by job type, average job value, and late jobs. One person builds it and the other three look at it. Prices are list prices as of November 2026, checked on Microsoft's pages except where noted.
| Set-up | Licences | Capacity | Monthly extra cost |
|---|---|---|---|
| Excel, no Copilot (existing Microsoft 365) | None extra | None | $0 |
| Excel with Copilot, builder only | One Copilot Business licence at $21 | None | $21 |
| Power BI Pro, no Copilot | Four Pro at $14 (builder plus three viewers) | None | $56 |
| Power BI Pro with Copilot | Four Pro at $14 | Fabric F2, about $263 left running | About $319 |
Microsoft's Copilot Business page lists $21 per user a month on an annual commitment, with a promotional $18 for the first year on offers made between 1 July and 31 December 2026; confirm the price on that page when you buy. Paying monthly on that annual commitment costs 5% more (about $22.05); the $25.20 month-to-month option comes from Microsoft's March 2026 partner pricing and is sold through Microsoft partners, not listed on that page (checked 20 November 2026). Enterprise Microsoft 365 Copilot is $30 per user on top of a base licence. Only the builder needs the Copilot licence: the three viewers open the workbook and read it.
The Fabric figure is Microsoft's launch price of about $0.36 an hour for F2 in one region, which depends on region. If you pause the capacity outside working hours it falls, as the sibling tutorial on Power BI Copilot cost and Fabric capacity works through. The point here is the shape: Excel with Copilot costs tens of dollars; Power BI with Copilot costs hundreds.
What can Copilot in Excel do for a dashboard?
Copilot in Excel can create and edit charts and PivotTables that stay linked to the source data, so it can draft most of a small dashboard from plain-English requests. Microsoft's support pages say it also answers questions about your data as charts, PivotTables, summaries, trends or outliers, and builds dashboards, planning tools and templates from scratch. It can import data from workbooks saved to OneDrive or SharePoint or on your computer. It needs a Microsoft 365 subscription that includes Copilot and an organisation that allows it.
Two retirements matter if you read older guides. The =COPILOT() worksheet function was retired on 14 September 2026, so any tutorial telling you to type it into a cell is out of date. The Copilot side pane in Excel remains, and that is what you use for dashboards.
A dashboard built this way is a sheet of PivotTables, charts and slicers. It is not a separate product. That is both its strength (you already know where to find it) and its limit (it is only as live as the workbook).
How do you build a small business dashboard in Excel with Copilot?
A small removals dashboard in Excel with Copilot comes from tidying the job log into a table, saving it to OneDrive or SharePoint, and asking Copilot for one chart at a time. The rest of this section is the main worked example, and it assumes the firm keeps a job log in a workbook with one row per job.
- Tidy the log first. One header row, no blank rows, no merged cells, one date per cell, and amounts as numbers rather than text. Format the range as an Excel table: Microsoft's support page says Copilot uses Excel tools such as tables, charts, PivotTables and formulas, and a named table gives it a clean range to work with.
- Save to OneDrive or SharePoint. Cloud-saved workbooks are what Copilot can read most reliably, and they let the three viewers open the same file.
- Open the Copilot pane and ask for one thing at a time. A prompt such as the one below works better than "make me a dashboard".
- Check each output against a number you know. If the log says 42 jobs last week and the chart says 41, find out why before adding the next visual.
- Add slicers. A slicer for month and job type lets viewers explore without breaking anything.
- Lock the layout. Protect the dashboard sheet so viewers cannot drag charts around.
Using the table JobLog, create a PivotTable of total revenue by job type
for the last 8 weeks, then a clustered column chart from it.
Put both on a new sheet called Dashboard. Do not change the JobLog table.
A realistic result is a PivotTable with five job types and a column chart, plus a short message from Copilot describing what it did. Illustratively: "Revenue by job type: house moves $18,400; flat moves $7,200; storage $3,100; office $5,600; man-and-van $2,900." Your task is to check that total against the log and fix any category that Copilot read oddly, for example "House move" and "House Move" counted separately because of a capital letter.
The step-by-step for a similar sales-view build, with the errors people meet, is in Create an Excel dashboard with Copilot: sales view errors and fixes.
If the spreadsheet side is the slow part, try The Complete Excel Course with AI Integration on Coding Liquids.
When is Power BI the better tool for a small business?
Power BI is the better tool when data must refresh unattended, many people need the same view, or the data outgrows a workbook. Microsoft's pricing page lists Pro at $14 per user a month, paid yearly, with the ability to publish and share reports, scheduled refresh of up to 8 times a day, a 1 GB model size and 10 GB of storage in the Pro column. The Free account can create reports but cannot publish or share them. Power BI Desktop is free to download.
These are the situations where the extra tool earns its place:
- Unattended refresh. A dashboard that must show this morning's bookings when the owner opens it at 8am, without anyone opening a workbook.
- Several data sources joined. Bookings system, card payments and the accounting package combined in one model with shared definitions of "revenue".
- Many readers, one definition. Twenty people looking at the same measure, not twenty copies of a workbook.
- Row volumes beyond comfortable Excel. A workbook with hundreds of thousands of rows becomes slow to open and risky to share.
- Row-level restrictions. Each branch manager sees only their branch.
If three or more of these apply, Power BI is probably the right tool whatever Copilot costs. If none apply, you are paying for infrastructure you do not need. For a profession-specific view, see Power BI for accountants: client dashboards and Copilot limits.
How do refresh and sharing differ between Excel and Power BI?
Excel refreshes when someone presses refresh or when a Power Query connection is set to refresh on open, while Power BI Pro refreshes on a schedule in the service. A dashboard that nobody trusts is worse than none. The usual reason is stale data, and the second is a file that exists in six versions.
Excel. The dashboard refreshes when the source data changes and someone presses refresh, or when the workbook is set to refresh on open for Power Query connections. If the data is typed straight into the workbook, there is nothing to refresh, and the person doing the typing is the refresh. Sharing is through a link to the one file on OneDrive or SharePoint, which keeps everyone on the same version if people are told not to download copies.
Power BI. Refresh is scheduled in the service on Pro. Microsoft's own documentation notes that scheduled refresh pauses after two months of inactivity and stops after four consecutive failures, so a dashboard that nobody opens can quietly freeze. Sharing needs a Pro licence on both sides in a smaller setup, which means the three viewers each cost $14.
A fair rule: if the person who builds the dashboard is also the person who feeds it, Excel is enough. If the feeding is automatic, think about Power BI. Where a flow could feed a workbook for free, Power Automate versus Power BI for getting sales numbers to a team compares the routes.
How long does a first dashboard take in Excel with Copilot or Power BI?
A first useful page takes about 1 to 2 hours in Excel with Copilot and 3 to 5 hours in Power BI, as illustrative planning figures for a non-specialist, not measurements. Adjust them to your own skill.
| Task | Excel with Copilot | Power BI |
|---|---|---|
| Tidy the data | 1 to 2 hours | 1 to 2 hours |
| First useful page | 1 to 2 hours | 3 to 5 hours (learning the tool) |
| Refresh set-up | Manual, minutes a week | 2 to 3 hours once |
| Sharing set-up | 15 minutes | 1 hour plus licences |
| Ongoing upkeep | Minutes a week | Minutes a week, plus fixing failed refreshes |
The first-page row is where most small firms decide. A person who knows PivotTables can produce a decent Excel page in an afternoon with Copilot's help. Learning Power BI's data model, relationships and measures takes longer, and the payoff only arrives when the data is large or live.
What should you ask Copilot for after the first chart?
After the first revenue chart, ask Copilot for the few numbers that answer the owner's questions, one request at a time, and check each result. A good dashboard answers a handful of questions, and the prompts should match them. The removals firm adds three more requests:
1. Add a card showing the number of jobs completed in the last 7 days.
2. Add a card showing average job value for the last 7 days.
3. Add a table of jobs marked Late this month with customer reference,
job type and days late, sorted by days late, largest first.
Illustrative output from the third request might be a six-row table with references such as "J-1042, house move, 2 days late". The things to fix are usually small: a date shown as a serial number instead of a date, a column that Copilot left in the wrong order, or a "days late" figure counted in calendar days when the firm measures working days. Tell Copilot the correction in plain words ("show days late as working days, Monday to Friday") and check the first three rows by hand.
Notice what Copilot has not done. It has not decided what "late" means for your business. That definition, in the log or in a note on the sheet, is the owner's job, and the dashboard is only as honest as that definition.
What layout fits a small dashboard on one screen?
A small dashboard fits on one screen at normal zoom when it follows a fixed order, left to right and top to bottom, because most small dashboards fail by being crowded. Use this order:
- Top row: three or four cards with the numbers the owner reads first (jobs, revenue, average value, late jobs), each with the period in the label.
- Middle: one trend chart over time, so a bad week is easy to spot.
- Lower left: one breakdown chart, such as revenue by job type.
- Lower right: the exceptions table, such as late jobs, because that is what gets acted on.
- Footer: source, last refreshed date and the reconciliation line.
The footer deserves a formula, not typing. A cell showing the latest date in the log (for example =MAX(JobLog[Date])) means viewers can see when the data was last updated, and a stale dashboard shows itself.
Can you start in Excel and move to Power BI later?
Yes, you can start in Excel and move to Power BI later, and the move stays cheap if you plan it from the start.
- Keep the data in a clean table with one row per event. Power BI imports that shape directly, so the work is not wasted.
- Write down your measures. "Revenue = sum of job price, excluding tax, by completion date." The same sentence becomes a measure in Power BI.
- Keep the dashboard sheet separate from the data sheet so you can retire the sheet without touching the data.
- Note the point at which you would move. For example: "More than six viewers, or the log passes 50,000 rows, or someone asks for the numbers before 8am."
When one of those triggers is hit, try Power BI Desktop, which is free to download, on the same table before anyone buys a licence. Publishing and sharing is where the $14 per user starts. Copilot is a separate question, covered in the sibling tutorial linked earlier, and for most small firms it is the last thing to buy rather than the first.
What the tools cost you in attention
Money is not the only bill. A dashboard asks for someone's attention every week. In Excel, that attention is spent pasting in new data and pressing refresh, a task that takes minutes and fails silently if the person is on holiday. In Power BI, attention goes to failed refreshes, expired credentials and licence admin; Microsoft notes that refresh stops after repeated failures. Neither is large, but both need a named owner and a backup, written down.
A short handover note works for either tool: where the data comes from, how often it updates, who updates it, what to do when a number looks wrong, and who to ask. Put it on the sheet or in the report description and review it each quarter.
Three businesses, three answers
An independent bookshop wants a monthly view of sales by category and a list of slow-moving stock. The till exports a spreadsheet each month. Excel with Copilot fits: one builder, monthly manual refresh, a page the owner prints for her accountant. Cost: $0 if she already has Copilot through her licence, or about $21 if she buys it.
A florist with web orders and a shop till wants to see daily takings combined. The two sources need joining, and the owner wants the view on her phone at 7am. This leans to Power BI, with the honest caveat that two connectors and a data model are more setup. She should ask first whether a single combined export would do.
A photography studio tracks bookings, deposits and delivered galleries. Three partners want a weekly page and a once-a-quarter review. A workbook with three PivotTables and a few slicers is enough, and the quarterly review can use a copy.
A mistake that makes Excel dashboards untrustworthy
An illustrative case: a removals firm's dashboard shows late jobs as 4 last week, and the owner is sure it was 7. The cause is a text column "Status" with "Late", "late " (with a trailing space) and "LATE" in different rows. The PivotTable counts them as three groups and the chart shows only "Late". The fix is a cleaned status column, a data-validation drop-down for future entries, and a reconciliation line on the dashboard: "Rows in log: 118. Rows counted: 118."
That reconciliation line is the cheapest trust-builder in any dashboard. If the two numbers differ, something was dropped. Power BI users need the same habit; the tool does not clean data for you.
If you want to pressure-test your choice before building, the questions below are quick to answer on paper.
Ten questions to settle it on one sheet of paper
- How many people need to see the dashboard, and how many will edit it?
- How often must the numbers update, and can a person do it?
- How many rows does the largest source have?
- Does the data live in one place or several?
- Does anyone need to see only part of the data?
- Do you already hold Microsoft 365 licences that include Copilot?
- Would a monthly export be good enough?
- Who will fix it when it breaks, and are they still here in a year?
- Do you already pay for Fabric or Premium capacity for another reason?
- What decision will the dashboard change? If none, do not build it.
If the answers are "few people, monthly, one source, small data, no restrictions", use Excel. If they are "many people, daily, several sources, restricted views", plan for Power BI and budget for capacity only if Copilot is a must-have. If you would like someone to run through these with you, that is part of what my AI implementation consultation covers; you can get in touch to see whether it fits.
Checking the finished dashboard works
Whichever tool you pick, run the same five checks before telling anyone it is live. Compare three headline numbers against the source by hand. Open it as a viewer would, on their device, and confirm they can see it. Change one row in the source and confirm the number moves. Read every chart title aloud to catch ambiguous labels such as "Revenue" with no period. Finally, write a two-line note on the dashboard saying where the data comes from and who to ask when it looks wrong. The note costs a minute and saves the first angry message.
Dashboard questions before you choose
Do I need Copilot to build an Excel dashboard?
No. PivotTables, charts and slicers do the job without any AI, and many small firms never need more. Copilot helps by drafting charts and PivotTables from a plain-English request, which saves time when you are not fluent in Excel. It requires a Microsoft 365 subscription that includes Copilot, so test the dashboard without it first.
Can Power BI Pro users use Copilot without extra purchases?
No. Microsoft's documentation says a Pro licence alone is not enough for Copilot in Power BI, which needs a paid Fabric capacity from F2 or Premium from P1. Pro still covers building and sharing normal reports without Copilot, at $14 per user a month.
Which tool refreshes automatically?
Power BI Pro supports scheduled refresh, up to 8 times a day according to Microsoft's pricing table. Excel can refresh Power Query connections when opened, and cloud-saved workbooks can be refreshed through other tools, but the setup is more manual. If a dashboard must update unattended every morning, that favours Power BI.
What if my data lives in a bookkeeping package, not a spreadsheet?
A monthly export into Excel is usually enough for a small firm. Check whether the package has a Power BI connector or exports to a spreadsheet on a schedule. If the data must refresh daily from a live system, Power BI's connectors are the stronger fit.
Is Excel with Copilot or Power BI better for a small business dashboard?
For most small firms, Excel with Copilot is cheaper and simpler: in the illustrative four-person example above it costs about $21 a month against about $319 for Power BI Pro with Copilot. Power BI is better when data must refresh unattended, many people need the same live view, or the data outgrows a workbook.
Further reads
Browse all AI-for-business tutorials
Sources: Microsoft Power BI pricing page, Microsoft Learn Copilot for Power BI overview, Microsoft Support pages on Copilot in Excel, checked November 2026.
