Same end goal, more control. The author builds the second list (Sessions 2026B) by hand-prompting one step at a time — choice columns first, then sample data via voice, then views, then a prettify pass.
Meta takeaways
Pace
Step-by-step prompts let you review each result before the next one.
Input
Voice is fast for long, list-shaped requests once the schema is set.
Polish
Styling is a separate prompt at the end, not part of the create step.
Sessions 2026B — built from prompts
The site from Guide 01 already exists. The moderate demo reuses it and asks SharePoint AI to create a second list called Sessions 2026B. The author types the columns this time — speakers, multi-line description, conference track with explicit choice values, and a status column with Approved / Rejected / Pending.
When you name the choice values yourself, the AI doesn't have to guess. The list comes back already shaped the way you want it.
This costs more keystrokes than the Agentic Builder, but every column matches what you asked for — no surprise additions.
The moderate tier isn't a step backwards. It's a deliberate trade — fewer surprises, more control over the schema.
Voice for the long parts
Once the list exists, the author switches to voice for the next prompts. Speaking is faster than typing once you're past the structured schema work and into long natural-language requests.
Voice shines when the prompt is a paragraph, not a column list. Use it after the schema is locked in.
The voice prompt that follows: generate 20 sample items for the list. You can assign them all to me as the speaker, but randomize the rest of the choices.
Sample data generation
That voice prompt generates 20 fake conference sessions — author as speaker on every row, but title, description, track, and status randomised. Useful test data for view design, with no manual data entry.
Asking the AI for sample data with anchors ("all assigned to me") and variation ("randomize the rest") gives you a dataset that looks real but is shaped how you need.
The pattern — pin some fields, randomise the rest — is reusable. Anywhere you want to exercise filters and groupings without hand-rolling data, the same prompt shape works.
Sample data isn't just for demos. View formatting is hard to design against an empty list — sample rows are how you actually see whether the styling works.
Views first, prettify second
Next, the author asks the AI to build two views — approved sessions and pending sessions. Then, separately, asks the AI to make the approved-sessions view prettier. Levels, status icons, emojis, basic colour treatment — applied as a follow-up step.
Views and styling are two prompts, not one. Get the data plumbing right first, then style.
The author notes the styling pass is a starting point — emojis and basic treatment that he can tweak. This is the same trade as Guide 01: AI brings you to 80%, you finish the last 20%.
A prettify prompt isn't applying the styling skills from Video 14 here. It's the built-in AI doing best-effort polishing — useful, but not brand-aware unless you load a style skill.
Flashcards — Applied
02 · Applied
Voice prompting
tap to reveal →
Speaking long natural-language prompts to SharePoint AI. Fast once the structured schema work is done and you're into paragraph-shaped requests.
← tap to flip back
02 · Applied
Sample data generation
tap to reveal →
Asking the AI to make N realistic-looking rows with some fields pinned and others randomised. Lets you design views against representative data.
← tap to flip back
02 · Applied
Pin-and-randomise
tap to reveal →
The sample-data prompt pattern: fix some fields ("assign all to me") and vary the rest ("randomize the choices") to get useful test rows quickly.
← tap to flip back
02 · Applied
View creation prompt
tap to reveal →
A dedicated prompt that builds a named filtered view ("show me approved sessions"). Runs after the list and data exist — separate from styling.
← tap to flip back
02 · Applied
Prettify pass
tap to reveal →
A follow-up "make this prettier" prompt that applies basic styling — levels, status icons, emojis. Not brand-aware unless a style skill is loaded.