.png)
.png)
3
Example Prompts & Workflows
The Headset MCP server works best when you ask the question you actually care about, not the query you think the database wants. This article collects prompts that work well, organized by what you're trying to get done. Copy them as-is, or use them as templates for your own.
Two habits make every prompt better:
- Name a time window. "Last 30 days", "this month", "June", "Q2". Without one, you're querying all-time data.
- Ask for the decision, not just the number. The AI can compare, rank, and summarize across multiple queries. "Which categories are growing and which are shrinking?" beats running two reports yourself.
For retailers
Daily and weekly check-ins
How were sales yesterday compared to the same day last week?
Give me a quick Monday recap: last week's revenue, units, and transactions vs. the week before, plus our top 5 products.
How does Saturday compare to the rest of the week over the last two months?
Product and category performance
What were our top 20 products by revenue in the last 30 days? Include units and margin.
Which categories grew and which shrank this month vs. last month?
Best-selling gummies in the last 90 days?
What's our flower sales breakdown by pack size this quarter?
Pricing, discounts, and margin
What's our average item price and discount rate by category for the last 30 days?
Which brands have the best gross margin, among brands where we sold at least meaningful volume?
Is our discount rate trending up or down week over week this quarter?
Inventory and purchasing
Which products have under two weeks of supply, sorted by how fast they're selling?
What's my total on-hand inventory value by category, and how does it compare to each category's trailing 28-day revenue?
Show me overstock: products with more than 12 weeks of supply and their on-hand value.
How many eighths do I have on hand across all stores, by brand?
Staff and operations
Who were our top budtenders by revenue last month?
Break down sales by order source for the last 30 days. How much is delivery vs. in-store?
What's our medical vs. recreational split by month this year?
Benchmarks
How did my store rank in the state last month across basket count, average item price, and revenue?
Where am I below the state median, and by how much?
Compare my stores against each other and against their state cohorts for last month.
Competitive intelligence (requires the add-on)
Which brands have the widest menu distribution in my state, and which of them am I not carrying?
Who are the 10 biggest competitor menus in my city, and how large are their menus vs. brand counts?
Which nearby stores carry [brand], and how widely is it distributed in my state?
For brands and distributors (Bridge)
Sell-through performance
How did our brand sell across all connected stores last month? Revenue, units, and how many stores sold it.
Which connected stores sold the most of our product line in the last 30 days, and which sold the least?
Show weekly sell-through for our top 5 SKUs over the last 12 weeks.
Retail execution and field sales
Which connected stores have under two weeks of supply of our products? I want a call list for my reps.
Which stores are currently stocked out of anything in our line?
Rank our connected stores by average daily velocity of our brand. Where is velocity dropping?
Assortment and market fit
What's our units-by-pack-size breakdown across connections? Are eighths or halves moving faster?
How does our med vs. rec split differ by state?
Which customer age groups buy our products, based on connected-store data?
Account reviews
Prepare a quarterly review for [store name]: our sell-through trend, share of their category sales, current on-hand, and weeks of supply.
Compare [store A] and [store B] on our brand's revenue, velocity, and stock position.
Multi-step workflows worth stealing
These are bigger asks that chain several tools together. They're where the connector earns its keep.
The weekly business review (retailer).
Put together my weekly business review: total sales vs. last week and the same week last year, top and bottom 5 products, category mix shifts, any products under 2 weeks of supply, and anything unusual you notice. Format it as a brief I can paste into Slack.
The reorder sanity check (retailer).
Before I place orders this week: which products are under 2 weeks of supply, what's their sales velocity trend over the last 4 weeks, and is anything on the low-supply list actually slowing down so I shouldn't over-order it?
The brand gap analysis (retailer, with Competitive Intelligence).
Compare my brand assortment to my market: my top 25 brands by revenue vs. the 25 most-distributed brands on menus in my state. Which brands are big on menus around me that I don't carry at all?
The stockout-prevention sweep (Bridge).
For every Active connection, find our SKUs with under 14 days of supply, then rank stores by how much revenue we'd lose per week if those items stock out, based on recent velocity. Give me the top 10 stores to contact today.
The new-market read (Bridge, with connected stores in multiple states).
Compare our brand's performance across states: revenue, units, store count, velocity per store, and trend over the last 6 months. Where are we strongest, and where is the biggest gap between store count and sales?
Prompting tips
Be specific about scope. "My Bellevue store", "across all connections", "Washington only". The AI can scope any question to stores, states, or cities.
Ask follow-ups. Every answer can be drilled into: "break that down by store", "now just for flower", "chart that weekly instead of monthly".
Ask for formats. "Format as a table", "give me a one-paragraph summary for my GM", "make it a bulleted brief for Slack". The data is the same; the packaging is up to you.
Let it resolve names. You don't need exact catalog spellings. Say "pre rolls", "eighths", or a rough brand name, and the AI looks up the canonical values in your data before filtering.
Sanity-check surprises. If a number looks off, ask "what filters and time window did you use for that?" The AI will tell you exactly how it scoped the query. For anything high-stakes, verify in Headset before acting.
