A SharePoint list with seven columns, ten rows of imported client data, pill formatting on the choice columns, and conditional formatting highlighting renewals due this month — all created from voice in under five minutes. The 5-minute challenge video shows the exact sequence.
This is what changes when Copilot in SharePoint is enabled on your site. The list settings UI, the column configuration screens, the view formatting JSON — none of them are how you build a list anymore. You describe what you want; the agent builds it.
The setup: an empty SharePoint site
The demo starts in an empty SharePoint site with no lists. The goal: build a client index list, populate it with data from an Excel spreadsheet, and apply formatting that makes the list readable.
Manually, this is 30 to 60 minutes of work. Create the list. Add the columns one at a time. Set the Status choice column options. Save. Find the Excel file. Open it. Copy data across with the column mapping working correctly. Open view formatting. Write JSON for the pill colours. Test. Write more JSON for the conditional formatting. Test again. Fix the JSON because the date function did not return what you expected.
With Copilot in SharePoint, the same job is one chat session with a few natural-language prompts.
The first prompt: create the list with columns
The video starts with the user opening the Copilot panel and dictating one long prompt that defines the entire list at once.
"In my SharePoint site, can you please create a list called client index? The title column should be called client name. I need columns for email address, postal address, contact name, status with the options for current and non-current, client type with options of vendor or supplier."
AI in SharePoint creates the list with all those columns in around fifteen seconds. The Status column is correctly set as a choice column with the two values. The Client Type column is correctly set as a choice column with the two values. The other columns are single line of text.
Notably the user did not need to specify the column types explicitly. AI in SharePoint inferred them from the wording: "options for X and Y" implies a choice column, "email address" implies text, and so on. For ambiguous columns you might want to be explicit ("contract renewal as a date column"), but the inference covers most common cases.
The list is created and the user is offered a link straight to it.
The second prompt: add a missing column mid-flow
Looking at the Excel data the user planned to import, one column is missing from the list: contract renewal date. The user asks to add it.
"Can you please add a date column called contract renewal?"
AI in SharePoint adds the column. The list now has all the columns needed for the Excel import.
This is the small thing that makes the pattern feel different. With the manual flow, the user would have to navigate back to the list settings, find Add Column, pick Date Time, name it, save, and return to where they were. The chat version is a single sentence.
The third prompt: import the Excel data
The user has a clients Excel file in the site's default documents library with ten rows of client data.
"Can you please add the clients that are listed in this spreadsheet to our new client index list?"
AI in SharePoint:
- Reads the Excel file to understand the data shape
- Checks the list schema to confirm column compatibility
- Identifies that the Contract Renewal column in the spreadsheet contains Excel serial date values
- Converts the serial dates to proper SharePoint date values
- Adds all ten rows to the list
About thirty seconds, ten records imported, dates converted automatically. The user typed seventeen words.
The fourth prompt: pill formatting on choice columns
With the list populated, the user wants the Status and Client Type choice columns to display as coloured pills rather than plain text.
"Can you make this list look a little bit prettier please? For our choice columns, let's look at adding some pills with different colours."
AI in SharePoint applies pill formatting:
- Status column: green pill for Current, red pill for Non-current
- Client Type column: blue pill for Vendor, purple pill for Supplier
No JSON written by the user. No view formatting screen opened. The output is the equivalent of about 40 lines of JSON manually authored.
The fifth prompt: conditional formatting
The final step is highlighting any client with a contract renewal date this month. This is the kind of view formatting that takes most users an afternoon of reading documentation to figure out.
"Can we just add some view formatting here please? Any client that has a contract renewal date this month, I want that client to be highlighted in red."
The first attempt fails. View formatting on dates is genuinely tricky and AI in SharePoint does not always land the right approach first time. The user asks it to try again.
"Can you try again?"
AI in SharePoint switches to a different approach using getMonth and getYear functions, which are more reliable than the complex date arithmetic of the first attempt. The row with a contract renewal date this month now shows in red.
This is worth noting because it sets the right expectation: the agent gets formatting right most of the time but not always on the first attempt. The retry pattern works because the agent has memory of the previous failed attempt and tries a different approach.
Total elapsed time
Under five minutes from the empty site to a fully built, populated, and formatted list. The video timeline shows the prompts and outputs in real time; there is no editing trick.
What changed for the user? Two things specifically.
The list settings UI never opened. Every action that would have required navigating to a list settings screen or a column configuration screen happened in chat instead. The settings UI still exists; the user just did not need it.
No JSON. View formatting on choice columns and conditional formatting on date columns both produced the equivalent JSON behind the scenes. The user never saw the JSON, never wrote any of it, and did not need to.
For the broader pattern beyond list creation, the three capabilities of Copilot in SharePoint explains where this fits (it is the "What to Produce" capability extended into list and view configuration).
When this matters most
Three scenarios where this saves the most time.
New site builds. Setting up a SharePoint site for a new team used to mean an afternoon of creating lists and configuring views. Voice cuts that to half an hour for most setups.
Form-style data capture. Teams that build internal forms (vendor requests, equipment loans, leave applications) using SharePoint lists get the structure plus the formatting in one session.
Existing-data ingestion. When data lives in a spreadsheet and needs to live in a SharePoint list, the voice + Excel import flow moves it across without manual copy-paste or PowerShell scripts.
For sites that will use Copilot in SharePoint regularly, putting a SharePoint.md context file in place first makes every list creation cleaner. The naming conventions, the choice column standards, the formatting preferences for your team are all stored once and applied every time. Without it, the list creation works but you have to re-specify the conventions in each prompt.
What to do this week
If your team has Copilot in SharePoint enabled on at least one site, two actions.
-
Try the 5-minute challenge yourself. Pick a small list you have been meaning to create. Open Copilot in SharePoint, dictate the list with columns. See how close you get in five minutes.
-
Put a SHAREPOINT.md context file in place so future list creation inherits your team's conventions. The template and the full walkthrough at the SharePoint.md article cover the pattern.
For a deeper look at the related capability of working with EXISTING list data through chat (querying, updating records, applying formatting), see AI in Microsoft Lists, the practitioner walkthrough.
If your team wants to run this pattern across multiple sites in a day rather than learning it incrementally, the Copilot in SharePoint Immersive is the format. We build lists, document libraries, and Skills in your tenant together so the team leaves with assets, not just demos.
