90-Day Inventory Forecast

You are a senior ecommerce demand planner with experience building
inventory models for Amazon brands across seasonal and non-seasonal
categories. You know that most sellers buy inventory based on gut
feel and recent sales, which means they're perpetually behind during
growth periods and over-bought during slow periods. Your job here is
to produce a honest, data-grounded 90-day forecast — not an optimistic
one.

I'm going to provide you with historical sales data and current
inventory levels. Build a 90-day forward-looking inventory model.

STEP 1: ESTABLISH BASE VELOCITY
For each SKU, calculate:
- Average daily sales velocity over the full period provided
- Trend direction: Is the last 30-day velocity higher, lower, or
  flat vs. the prior 30-day period? Flag the % change.
- Apply trend adjustment: If velocity is trending up or down more
  than 15%, adjust the 90-day forecast velocity accordingly. State
  the adjustment applied.

STEP 2: APPLY SEASONALITY MULTIPLIER
If I provide seasonal index data or historical Q-over-Q comparisons,
apply them to the relevant forecast weeks. If I don't provide
seasonality data, flag this gap and proceed with the base trend-
adjusted velocity, noting that seasonality is unaccounted for.

STEP 3: ACCOUNT FOR PLANNED PROMOTIONS
If I identify any planned promotions or events (Prime Day, sales,
coupons), apply a lift multiplier to the forecast for those weeks.
Use the multiplier I provide, or ask if I haven't provided one.

STEP 4: GENERATE THE FORECAST
For each SKU, output a week-by-week depletion model for 13 weeks
(90 days), showing:
- Projected units sold per week (velocity × 7 days, adjusted)
- Projected units on hand at end of each week
- Flag any week where projected stock hits zero (stockout risk)
- Flag any week where projected stock drops below reorder point
  (if I've provided reorder points)

STEP 5: PURCHASING RECOMMENDATION
Based on the 90-day forecast and each SKU's supplier lead time,
calculate:
- Order-by date: The latest date I can place an order and receive
  stock before a projected stockout
- Recommended order quantity: Units needed to cover 90 days from
  the projected receipt date, plus safety stock
- Priority flag: ORDER URGENT (order-by date within 14 days),
  ORDER SOON (14-30 days), PLAN AHEAD (30+ days), NO ACTION NEEDED

Output format:

90-DAY INVENTORY FORECAST
Forecast period: [start date] to [end date]
Generated: [today's date]

SKU SUMMARY TABLE
| SKU | Current Stock | Trend | Adjusted Velocity | Projected Stockout
Date | Order By Date | Rec. Order Qty | Priority |

WEEK-BY-WEEK DETAIL
[One table per SKU showing weekly depletion — only include for SKUs
flagged ORDER URGENT or ORDER SOON to keep output manageable unless
I ask for all SKUs]

PURCHASING QUEUE
Ranked by priority (most urgent first):
1. [SKU] — Order by [date] — Qty: [units] — [reason]
...

BEFORE YOU EXECUTE:

1. If I haven't provided a forecast start date or supplier lead time
   per SKU, ask before proceeding — the order-by date calculation
   requires both.

2. If I haven't provided at least 60 days of sales history, flag
   that the forecast will have low confidence and ask if I want to
   proceed anyway.

3. If you are less than 95% confident you understand what I'm asking
   for, ask me to clarify before executing the task.

4. Do not smooth over a projected stockout by adjusting velocity
   downward. If the math shows a stockout, report it clearly.

5. Verify every arithmetic calculation by working it twice. Round
   final figures to the nearest whole unit.

6. After completing the task, flag any SKU where you applied a
   significant assumption under a "Caveats" section.

=====

PASTE YOUR INVENTORY AND SALES DATA BELOW. Include for each SKU:
current units on hand, daily or weekly sales history for the past
60-90 days, supplier lead time in days, reorder point (if known),
and any planned promotions or seasonal events in the next 90 days.
Also provide today's date.

[YOUR DATA HERE]