BidX Amazon Advertising Blog | BidX

Talk to Your Amazon Ads Data: Introducing the BidX MCP

Written by Max Hofmann | Wed, Jul 15, '26

What if you could ask your Amazon advertising data questions in plain English and get answers in seconds? Not dashboards. Not exports. Just: "How did Prime Day go compared to last year?" and a complete analysis, charts included.

That is exactly what the BidX MCP makes possible. As an Amazon Ads MCP it connects AI assistants like Claude directly to your BidX account, turning every campaign, keyword, ASIN, and competitor data point into something you can simply have a conversation with.

What is MCP, and why should advertisers care?

MCP (Model Context Protocol) is an open standard that lets AI assistants securely connect to external tools and data sources. Think of it as a universal adapter: instead of copy-pasting CSV exports into a chatbot, the AI queries your live data directly, runs the analysis, and explains what it found.

For Amazon advertisers, this MCP changes the daily workflow fundamentally. Questions that used to mean twenty minutes of clicking through reports, filtering date ranges, and building pivot tables now take one sentence.

How to connect

The BidX MCP is currently in beta (invite only). Once you have access, setup takes under a minute:

  1. Get your access - reach out to your BidX account manager to be enrolled in the beta.
  2. Add the connector - in Claude (or any MCP-compatible AI client), add a new custom connector and point it to the BidX MCP endpoint: https://app.bidx.io/mcp/sql
  3. Authenticate - you will be redirected to log in with your BidX credentials. Access is scoped to the advertising profiles your BidX account can already see, nothing more.
  4. Start asking - that is it. No API keys to manage, no data pipelines to build.


See the videos below for your reference.

 

 

The MCP exposes two tools to the AI:

  • discover-sql-schema - lets the assistant explore which tables and columns are available, so it writes correct queries on its own.
  • run-sql-query - executes read-only SQL against your data and returns the results.

Read-only is the key phrase: the MCP can analyze everything but change nothing. Your bids, budgets, and campaigns stay untouched.

What data can you query?

The BidX MCP is not limited to ad reports. It brings together the full picture of your Amazon business in one queryable layer:

Advertising performance Campaign, ad group, product ad, and keyword/target level performance across Sponsored Products, with daily granularity. Spend, sales, clicks, impressions, conversions, ACoS, ROAS - sliceable by any dimension and any date range.

Campaign structure Your full account structure: campaigns (including auto vs. manual targeting type), ad groups, product ads mapped to ASINs, and every keyword and product target with its match type and current bid.

Total sales (paid + organic) This is where it gets interesting. The MCP includes total performance data per ASIN - combined paid and organic sales and orders. That means you can measure the organic halo of your advertising, not just the attributed sales in the ads console.

Search Query Performance (SQP) Amazon's weekly SQP data, showing how your ASINs perform across the full search funnel: impressions, clicks, add-to-carts, and purchases per search query, including your share versus the market total.

Keepa market intelligence Integrated Keepa data adds the competitive dimension: current snapshots per ASIN (price, rating, review count, bestseller rank, estimated monthly units), historical BSR time series, and automatically built competitor sets per ASIN based on Amazon's own category bestseller lists.

Real example cases

These are workflows we run daily - each starts with nothing more than a sentence in a chat window.

1. The Prime Day post-mortem "Analyze our Prime Day performance vs. a 30-day baseline and vs. last year." The AI detects the exact event dates by finding the revenue spike in daily data (no hard-coded assumptions), pulls headline uplift, day-by-day ROAS curves, keyword-level winners, and the organic halo per ASIN - then builds a recap deck or interactive dashboard on top. What used to be a full day of analyst work now takes a coffee break. In one recent review, the standout finding was an ASIN with zero ad spend that still saw a significant organic lift during the event - a clear candidate for paid budget at the next deal event, and exactly the kind of insight that never surfaces in a standard ads report.

2. Competitor benchmarking in seconds "Who are the top 3 competitors for this ASIN, and how do we compare?" Using the Keepa competitor sets, the AI pulls each competitor's price, rating, review count, and bestseller rank, then positions your product against them. In one case this immediately revealed that the review-count gap, not price or rating, was the biggest lever holding back category rank.

3. Portfolio health check via BSR trends "Show me 30-day BSR trends for our top 10 ASINs." Historical Keepa data surfaces momentum instantly: which products are climbing the rankings, which are quietly deteriorating. One deteriorating ASIN we caught this way would have been invisible in the ads dashboard, because its ad metrics looked fine - the organic slide only showed up in BSR.

4. Search funnel diagnostics with SQP "Where are we losing shoppers in the search funnel for our best-seller?" SQP data shows, query by query, whether the problem is impression share, click-through, add-to-cart, or purchase conversion - and therefore whether the fix is bids, main images, pricing, or listing content.

5. Bid efficiency audits "Plot our bids against actual CPCs for all active keywords." A scatter of bid vs. realized CPC exposes over-bidding pockets and keywords where there is headroom to push - across thousands of keywords, in one query.

Why this beats exporting CSVs into a chatbot

You could always download reports and upload them to an AI. The MCP approach is different in three important ways:

  • Live and complete. The AI queries current data across all tables at once, including joins a manual export cannot easily produce (ad performance joined to organic sales joined to Keepa rank).
  • Iterative. Every answer invites a follow-up. "Now break that down by ASIN." "Compare it to May." "Which keywords drove that?" No new export needed.
  • Reusable. Recurring analyses can be captured as skills - reusable playbooks the AI follows step by step. Our Prime Day review skill turns a complex multi-query analysis into a one-line command.

Getting started

The BidX MCP as well as this blog article will be improved on an ongoing basis. If you want your Amazon ads data to answer back, reach out to us at BidX account manager for access - and bring your hardest question. That is usually the best demo.