Skip to main content
Nova Analytics
Try for free

How can AI help an Amazon Vendor Central account manager?

AI can help an Amazon Vendor Central account manager by running the weekly account review, and for that the assistant needs live access to your sell-in, Amazon's sell-out, stock, purchase orders and payments. Nova's MCP connector gives Claude, ChatGPT and other assistants that access to Vendor Central. Through Nova, the assistant compares sell-in and sell-out week on week, names the products Amazon should be ordering, lists past due purchase orders and disputable invoices, and prepares the points to raise with your vendor manager.

Updated

Connect your AI Copy the playbook

The playbook

Every step calls a Nova tool that reads your live Amazon data, so paste this into Claude, ChatGPT or any assistant connected to Nova's MCP server. Run it on Monday, or before a call with your Amazon vendor manager.

This needs an assistant connected to Nova. Start the 14-day free trial, then connect your assistant.

Use Nova to run my weekly Amazon Vendor Central account review and tell me what to act on first.

Period: the last complete week against the week before. If I name other dates, use those.

Scope: my Vendor Central accounts only. If I hold several Vendor codes, say which code each figure comes from.

1. List the accounts. Call Nova's list_accounts with channel "VENDOR". Give each account's marketplaces, Vendor codes and brands.

2. Read sell-in and sell-out. Call Nova's query_metrics with channels ["VENDOR"], metrics ["sales", "units", "vendorShippedRevenue", "vendorShippedUnits", "netProfit"], the period and the comparison period. Sell-in is what Amazon bought from you, sell-out is what Amazon shipped to its customers. Say which basis the sales figure uses, as `channelNote` states it.

3. Find the products that moved. Call Nova's find_movers with channels ["VENDOR"], metric "vendorShippedUnits", dimension "childAsin", direction "down", then direction "up". Name the five largest in each direction.

4. Check Amazon's stock. Call Nova's query_vendor_inventory with states ["should_order", "oos_po_open", "low_cover"], sorted by units4Weeks. For each product give the state, sellable units, weeks of cover and open purchase order units. Say the snapshot date: it is the end of the last week Amazon reported.

5. Check what Amazon loses money on. Call Nova's query_vendor_inventory with negativeNetPpm true. These are products where Amazon's Net PPM was below zero last month, which often comes before Amazon stops ordering.

6. Read the purchase orders. Call Nova's query_vendor_purchase_orders with status "open" and level "order". From `position`, give what is open, due this week and past due. From `rates`, give confirmation rate, fill rate and lead time.

7. Read the money. Call Nova's query_vendor_invoices with range "open". From `balance`, give what Amazon owes and what is short-paid. Then call it with bucket "disputable" and list the invoices whose 90-day window closes soonest. Then call Nova's query_vendor_remittances with level "line" and class "deductions" for the last 4 weeks and give deductions by family.

8. Check the listings. Call Nova's query_listing_health with channels ["VENDOR"], bands [0, 1], sort "severity", limit 10. Name any suppressed listing among the products from steps 3 and 4.

9. Report. Write one paragraph on the week, a table of products to act on with the reason, and three points to raise with my Amazon vendor manager.

Rules: never add amounts of different currencies, and never add 1P and 3P figures yourself. Rates from query_vendor_inventory are fractions (0.93 = 93 %) and rates from query_vendor_purchase_orders are percentage points, so convert before you compare them. A forecast that returns null means Amazon gives no forecast, which is not zero. Sessions and conversion are not reported on Vendor Central: say so when asked. Co-op is money you agreed to pay, so never present it as money Amazon owes you. Relay every notice Nova returns. You cannot confirm purchase orders, file disputes or change anything in Vendor Central.

Raw markdown

How Nova customers use this

Real requests from Nova customers, names removed.

  • “Review our Vendor Central account over the year: sales, Amazon's demand, profitability, purchase orders, stock, deductions and listings.”

    What the assistant did

    It read sell-in, Amazon's shelf demand and profit, then purchase order volume with acceptance and fill trends, stock cover and Net PPM, payments and deductions by category, open and short-paid invoices, and the listings with a problem.

    Nova tools used (8)
    1. list_accounts
    2. list_metrics
    3. query_metrics
    4. query_vendor_purchase_orders
    5. query_vendor_inventory
    6. query_vendor_remittances
    7. query_vendor_invoices
    8. query_listing_health

  • “Build the numbers for our management deck: sell-out this year against last, Net PPM by quarter and the sell-in outlook for the last quarter.”

    What the assistant did

    It compared year-to-date sell-out with last year by account, by month and by quarter, ranked the products that lost the most, read Net PPM on the top products, and totalled open purchase order value to support the sell-in outlook.

    Nova tools used (4)
    1. query_metrics
    2. find_movers
    3. query_vendor_inventory
    4. query_vendor_purchase_orders

  • “Several products are down on last year. When did each one start falling, and was it stock, price or something else?”

    What the assistant did

    It laid out monthly sell-out and stock for each product, then read weeks of cover, confirmation and fill rates, past due units, cost price by quarter and Best Sellers Rank, and compared them with sibling products in the same families.

    Nova tools used (5)
    1. list_accounts
    2. list_metrics
    3. query_metrics
    4. query_vendor_inventory
    5. query_bsr

  • “Weekly check: which products does Amazon lose money on, where are best sellers running out, and what was deducted this week?”

    What the assistant did

    It read monthly Net PPM and shelf profit, listed the products with negative Net PPM that Amazon may stop ordering, checked that product costs were entered, flagged best sellers out of stock, and totalled the past week's deductions.

    Nova tools used (5)
    1. list_metrics
    2. query_metrics
    3. query_vendor_inventory
    4. get_setup
    5. query_vendor_remittances

Which parts of the job can the assistant take on?

Weekly account review

What the assistant reads

Sell-in, sell-out, Net PPM, stock, purchase orders, payments

What you get

What moved, why, and what to act on first

Vendor manager call prep

What the assistant reads

Products Amazon should order, past due lines, negative Net PPM

What you get

Three points to raise, with the figures behind them

Purchase order follow-up

What the assistant reads

Open, due and past due orders, confirmation and fill rates

What you get

The lines to ship first and the products that drag rates down

Deductions and shortages

What the assistant reads

Payments, deductions by family, short-paid invoices

What you get

What Amazon took, what it owes, what you can still dispute

Product decline diagnosis

What the assistant reads

Monthly sell-out, stock, weeks of cover, cost price, Best Sellers Rank

What you get

When each product started falling and the likely cause

Management deck

What the assistant reads

Sell-out this year against last, Net PPM by quarter, open order value

What you get

The numbers for the slides, by account, month and quarter

Two jobs have their own playbook: the purchase order playbook for fill rate and what ships next, and the deductions playbook for what Amazon deducted and still owes you.

How to read the result

Sell-out up, sell-in flat

What it usually means

Amazon is selling through its stock without reordering

What to do

Check weeks of cover and raise the products in should_order

Sell-in up, sell-out flat

What it usually means

Amazon is building stock

What to do

Watch for excess and aged units in the coming weeks

Out of stock with a purchase order open

What it usually means

Units are confirmed and have not arrived

What to do

Ship the past due lines first

Negative Net PPM on a product

What it usually means

Amazon loses money on each unit

What to do

Review cost price and terms on that product before Amazon stops ordering it

Low confirmation rate

What it usually means

You confirm fewer units than Amazon orders

What to do

Check stock and production before the next PO wave

Deductions rising faster than invoices

What it usually means

Shortage claims, chargebacks or co-op are growing

What to do

Open the largest family and check what is still disputable

Limits

  • Amazon's stock is the last weekly snapshot Amazon reported, a Saturday, never today.
  • Amazon does not report sessions or conversion on Vendor Central, so the assistant cannot explain a drop through traffic.
  • Reviews, Search Query Performance and Subscribe and Save are not read for Vendor Central.
  • Purchase orders, invoices and payments stay in each document's own currency.
  • Nova reads Vendor Central. It does not confirm purchase orders, file disputes or edit listings there.

How do I connect Vendor Central to my assistant?

Connect your Vendor Central account to Nova, then add Nova's MCP server to your assistant. The guides cover Claude and ChatGPT. The Vendor Central MCP server page lists what the assistant can read.

Questions sellers ask

Can the assistant confirm purchase orders or file disputes?

No. Through Nova the assistant reads Vendor Central and prepares the work: which lines to confirm or ship, which invoices to dispute and before which date. You act in Vendor Central yourself.

Does it work with several Vendor codes?

Yes. Each Vendor Central account in Nova carries its Vendor codes and brands, and the playbook asks the assistant to say which code each figure comes from.

Can it read 1P and 3P together?

Yes. Nova keeps 1P and 3P accounts in one workspace with a 1P / 3P filter, so the assistant can read both. It never adds 1P and 3P figures by hand.