Amazon MCP Campaign Creation Guide: How to Build Sponsored Products Campaigns in Natural Language

Amazon

September 09, 2026

TL;DR

Amazon Ads MCP Server includes an end-to-end Sponsored Products campaign creation workflow that can turn a natural-language instruction into the API operations required to create a campaign.

The end-to-end Sponsored Products campaign creation workflow can orchestrate multiple steps required for campaign creation through a single natural-language request. The information required depends on the Sponsored Products campaign structure being created. Inputs can include the advertised product, campaign name, budget, targeting type, bids, and, for manual targeting, keywords or product targets.

Sponsored Products supports automatic and manual targeting. Manual keyword targeting can use broad, phrase, and exact match types, while product targeting is also available.

A practical workflow is to set the advertising profile, describe the campaign, ask the AI to present the proposed configuration for review, and confirm before requesting execution.

MCP cannot decide your strategy for you: the ASIN choice, the bid level relative to your margin, and the keyword selections are human decisions. The AI executes; you set the direction.

Creating a Sponsored Products campaign manually takes about 15 to 20 minutes if you know what you are doing. You log into Campaign Manager. You create the campaign. You add an ad group inside it. You add the product. You set targeting. You save. You realise you forgot to set the placement modifier. You go back.

That is not a particularly hard process. But when you are launching five campaigns across three ASINs before a promotional event, it is an entirely real way to spend two hours: clicking through the same sequence repeatedly, checking each setting as you go.

Amazon Ads MCP can use a natural-language instruction to initiate an end-to-end Sponsored Products campaign creation workflow. If you want to review the proposed setup before execution, explicitly ask the AI to show the campaign details and wait for confirmation. What previously required three or more separate operations happens in one reviewed workflow.

That time saving is not the most interesting part. The most interesting part is what it means for teams that are not PPC specialists: authorized users can use natural-language instructions through a compatible AI agent to create supported Sponsored Products campaigns without manually navigating each campaign-creation step in Campaign Manager.

Campaign creation is where the time savings are most immediately visible with Amazon MCP. The strategic thinking still has to come from the person running the account. The execution is where the AI earns its place.

— Meher Patel, Founder & CEO of Hector AI and Amazon Ads Top 20 Global Partner

What Amazon MCP Actually Does When You Create a Campaign

The Old Way: Three Operations, Three Screens

Creating a Sponsored Products campaign directly through the API can require multiple related operations for campaign entities, ad groups, advertised products, and targeting, depending on the campaign structure. First, you created the campaign and defined its budget and bidding strategy. Second, you created the ad group inside the campaign and set the default bid. Depending on the campaign structure, you create advertised product entities within the ad group and configure the applicable keyword or product targets.

Each step had to succeed before the next could begin. If you were building this programmatically, you would be writing the orchestration logic yourself: handling the sequence, managing the IDs returned at each step, and ensuring each component pointed to the right parent.

The New Way: One Instruction, One Reviewed Package

Amazon built its MCP campaign creation tool as an orchestrated workflow, not a raw API wrapper. When you give it a campaign creation instruction, the workflow can orchestrate the required campaign-creation steps. Users should explicitly request a review of the proposed campaign settings before authorizing execution, where their AI client and workflow support this behavior.

Think of it like placing a catering order. You don't specify each dish, the order it should be prepared, the plating sequence, or the timing. You say: "I need a lunch for 12 people, Mediterranean theme, $20 per head, ready at noon." The caterer handles the sequence. You review the menu before confirming.

That is what the Amazon MCP campaign creation tool does. You describe the desired campaign in natural language, and the MCP workflow can coordinate the corresponding Amazon Ads API operations. For additional control, ask the AI to present the proposed settings and wait for confirmation before execution.

A single prompt replaces 15-20 minutes of manual Campaign Manager setup (Amazon Ads official announcement, February 2026). The time difference is significant. The orchestrated workflow can reduce repetitive manual setup, but advertisers should still validate campaign settings before execution.

The Three Sponsored Products Campaign Structures MCP Supports

Auto Campaigns: Let Amazon Find the Searches

With Sponsored Products automatic targeting, Amazon matches your ads to relevant shopping queries and products based on information about your advertised product. With automatic targeting, Amazon automatically matches your ads to relevant shopping queries and products. Advertisers can set bids and can also manage bids across automatic targeting groups.

Auto campaigns are most useful when you are launching a new product and do not yet have search term data to guide manual targeting. They generate the search term report that tells you which queries are converting, which you then harvest into manual campaigns.

Exact Match Campaigns: Target What You Already Know Converts

An exact match campaign shows your ad only when a shopper's search query is an exact match (or close variation) of a keyword you specify. Exact match provides tighter control over which shopper queries can match a keyword, but performance and ACoS depend on bids, competition, conversion rate, product relevance, and other factors.

Exact-match keyword targeting can be useful when advertisers have identified relevant keywords they want to target precisely, whether from historical search-term data, keyword research, or other sources.

Broad Match Campaigns: Cast a Wider Net With Discipline

Broad match provides the widest keyword match coverage and can match shopper queries containing variations, related terms, synonyms, and other relevant variations of the keyword. It reaches more searches than an exact match but requires active negative keyword management to prevent budget waste on irrelevant terms.

Broad match is the least restrictive keyword match type within manual keyword targeting. Automatic targeting is a separate Sponsored Products targeting method. They are useful for growing markets, seasonal events, or categories where purchase language is not yet fully mapped.

Before You Type Anything: The Five Inputs You Need Ready

The inputs required to create a Sponsored Products campaign depend on the targeting and campaign structure requested. The AI agent may request additional information when required inputs are missing. Having them ready before you start prevents the AI from having to ask follow-up questions mid-prompt.

Input	What It Is	Example
ASIN	The product you are advertising	B0ABC1234X
Campaign type	Auto, exact, broad, or a combination	Exact match manual
Daily budget	The average amount you are willing to spend on the campaign per day	$40 per day
Starting bid	The initial keyword or target bid	$0.85 starting bid
Campaign name	A name you will recognise in Campaign Manager	SP-Exact-BlenderV2-US-Jul26

For keyword-targeted campaigns (exact or broad), you will also need a keyword list. If supported by the connected tools or AI client, you can ask for keyword suggestions. Review suggested keywords before using them as campaign targets. The cleaner the inputs, the cleaner the output the AI prepares.

Creating a Campaign With Amazon MCP: The Complete Walkthrough

Step 1 · Set Your Active Advertising Profile
Before any campaign creation prompt, confirm which advertising profile the MCP connection is pointed at. A wrong profile means a campaign lands in the wrong account.
Type: "List my available advertising profiles and confirm which one is currently active."
If the connected MCP tools expose advertising-profile discovery and selection, ask the AI to list the advertising profiles available to your authorized account and verify the correct account and marketplace before creating a campaign.
Watch out: Never skip this step. Creating a campaign under the wrong advertising profile can result in changes being made to the wrong account, so verify the account context before executing write actions.

Step 2 · Tell the AI What You Want to Build
In one prompt, describe the campaign you want. Include all five inputs from the table above. The more specific you are, the more complete the package the AI prepares.
If you want the AI to suggest keywords, tell it so. If you want to supply your own keyword list, paste it directly in the prompt. You can ask the AI agent to create multiple campaign structures, but whether they can be handled in a single workflow depends on the tools and orchestration supported by the connected Amazon Ads MCP implementation.
You do not need to use technical language. Plain English works because the MCP tool handles the translation into structured API calls.

Step 3 · Review the Package the AI Prepares
Ask the AI agent to present the proposed campaign structure and wait for your confirmation before executing the campaign creation workflow. This is the IKEA flat-pack moment: all the pieces are laid out in front of you before you start assembling. You check that everything is correct before anything is built.
Check: the campaign name matches your naming convention, the daily budget is correct, the starting bid is right for the category, the ASIN is the right one, and (if applicable) the keyword list contains what you intended.
This review step is where you can also catch issues the AI cannot see, like the fact that you have already launched this ASIN in another campaign at a similar bid, and you need to differentiate them.
Watch out: Do not skip the review because a prompt 'looks right.' The review is what separates an AI-assisted workflow from an automated one.

Step 4 · Confirm and Launch
Once the review is complete and you are satisfied with the package, confirm the launch. After execution, review the tool response to verify which campaign entities were successfully created and capture the identifiers returned by the workflow.
Check Campaign Manager within a few minutes to confirm the campaign is showing as active. Verify the budget, bid, and ASIN match what you specified. If the campaign settings match your request, verify the campaign's current status and eligibility in Campaign Manager. Campaign creation does not necessarily mean the ads are immediately serving.

Copy-Ready Campaign Creation Prompts
These prompts are written to include all five inputs and a confirmation step. Adjust the ASIN, budget, bid, and keywords to match your account before using.

Auto Campaign: Single ASIN Launch
"Create a Sponsored Products auto campaign for ASIN B0ABC1234X. Name it SP-Auto-BlenderV2-US. Daily budget $35, starting bid $0.75. Show me the full campaign structure before launching."
Note: The auto campaign type is implied by 'auto campaign.' The confirmation step is explicit. Ask the AI to return the proposed campaign settings, targeting, advertised products, and other relevant configurations for review before requesting execution.

Exact Match Campaign: Known Keywords
"Create a Sponsored Products exact match manual campaign for ASIN B0ABC1234X targeting these keywords: protein blender, portable blender bottle, gym blender cup. Name it SP-Exact-BlenderV2-US. Daily budget $50, starting bid $0.90 per keyword. Confirm the structure with me before launching."
Note: Pasting the keyword list directly in the prompt means the AI does not need to generate suggestions: it builds the ad group around your list. The 'confirm structure with me before launching' phrase is the confirmation step.

Multi-ASIN Launch: Three ASINs in One Pass
"Create one auto Sponsored Products campaign for each of these ASINs: B0ABC1234X, B0DEF5678Y, B0GHI9012Z. Name them SP-Auto-[ASIN]-US. Daily budget $25 each, starting bid $0.70. Show me all three campaign structures before creating any of them."
Note: Asking to see all three structures before creating any of them is critical here. It ensures you review the full batch before committing. Ask the AI to display the proposed configuration for all three campaigns before requesting creation. The exact presentation and orchestration depend on the AI client and available MCP tools.

Marketplace Expansion: Replicate to a New Country
"Replicate campaign SP-Exact-BlenderV2-US to the UK marketplace. Show me what will be replicated, flag any settings that need to be adjusted for the UK, and wait for my confirmation before creating the UK version."
Note: Amazon Ads MCP Server includes workflows that can help expand supported campaigns to additional locales. Review which campaign settings, products, targeting, and other elements are supported in the destination marketplace before execution. Asking it to 'flag settings that need adjustment for the UK' prompts the AI to surface currency, pricing, and eligibility considerations you might not think to check unprompted.

What to Review Before You Confirm the Launch
The AI's review package will show you the full campaign structure. Here is a checklist of what to verify before typing 'confirm':

Pre-Launch Review Checklist
Campaign name matches your naming convention: you will need to find this in Campaign Manager later.
Daily budget is correct: not a default figure the AI assumed.
Starting bid is appropriate for your category: too low and your ads will not show; too high and you burn budget before getting data.
ASIN is the correct one: verify that the product title the AI returns matches the product you intended.
Campaign type is correct: auto vs manual, and if manual, that the match type is what you specified.
For keyword campaigns, the keyword list contains your intended terms and no accidental duplicates.
Review existing campaigns for overlapping targeting and determine whether the new campaign serves a distinct strategic purpose.

What MCP Cannot Decide for You?
Amazon Ads MCP Server connects AI agents to Amazon Ads API functionality and can support both analytical and execution-oriented advertising workflows. Strategic decisions should still incorporate advertiser objectives and business context. Four decisions come from you before the AI can create a campaign that is actually worth running.
Which ASIN to advertise: The AI may not have access to all product-readiness information required to determine whether an ASIN should be advertised. Review factors such as listing quality, ratings, inventory, pricing, and business objectives separately. Product readiness is a human decision.
What bid to set relative to your margin: The AI can accept a bid you give it. Amazon Ads MCP should not be assumed to have access to proprietary business inputs such as product costs, margins, or your internally calculated break-even ACoS unless those inputs are provided through another connected system. Bid decisions require margin context that the AI does not have.
Which keywords to target: The AI can suggest keywords based on the product or generate them from your listing content. Historical search-term performance may be available through Amazon Ads reporting capabilities. Use available performance data when selecting or evaluating keywords rather than relying solely on AI-generated suggestions. Your search term report data is the most valuable input here.
When to pause or scale: Once the campaign is live, performance decisions: when to pause a keyword, when to increase a bid, when to add negatives: require data the AI can pull, but judgement you must apply. Platforms like Hector Ai can automate these ongoing decisions through rule-based automation, but the rules still need to be set by someone who understands the business.

What to Do After Launch
Campaign creation is the beginning, not the end. Think of it like planting a seed: the planting takes minutes; the growing takes weeks, and the tending determines whether it produces anything useful.

Post-Launch Monitoring Timeline
Day 1-3: Check that the campaign is spending and ads are showing. If a newly launched campaign receives little or no delivery, investigate factors such as bid competitiveness, budget, targeting, product eligibility, Featured Offer status where applicable, relevance, and campaign status.
Review search-term data once the campaign has accumulated enough meaningful traffic and conversion data to support decisions. The appropriate review period varies by traffic volume, category, budget, and conversion cycle. Look for converting terms to harvest into exact match campaigns, and irrelevant terms to add as negatives.
Evaluate ACoS against your business target once sufficient attributed performance data has accumulated, accounting for the applicable attribution window and your account's traffic and conversion volume. For exact match campaigns, review which keywords are generating orders and which are consuming budget without conversion.
Monthly: Check for cannibalisation: are multiple campaigns competing against each other for the same search terms? Clear campaign structures can make targeting, budgeting, optimization, and performance analysis easier, but campaign structure alone does not guarantee better performance.

Through Amazon MCP, you can run the post-launch review as a natural language query: "Show me the search term report for campaign SP-Auto-BlenderV2-US for the last 14 days, sorted by spend, with conversion rate for each term." The async reporting window means you will wait 10-30 minutes for the result, but the data comes back structured and ready to act on.

Frequently Asked Question

End-to-end means Amazon MCP's campaign creation tool handles the full multi-step sequence: creating the campaign entity, creating the ad group inside it, and creating the ads within the ad group, as a single orchestrated operation triggered by one natural-language instruction. Before MCP, each of these was a separate API operation that had to be executed and confirmed in sequence. End-to-end means you describe the outcome once, and the tool handles the sequence.

Yes. The campaign workflow must ultimately identify the product to advertise. You can provide the ASIN directly, or the connected tools may be able to identify products from other available product information or account context. If you are unsure of the ASIN, check your product listing URL in Seller Central: the ASIN appears in the URL and on the product detail page. If the connected MCP tools expose product or advertised-product discovery capabilities, you may be able to ask the AI to identify products available for advertising. This should not be described as requiring Seller Central data specifically.

Yes. Amazon Ads MCP Server exposes multiple Amazon Ads capabilities through its available tools. Avoid relying on specific tool prefixes in the article unless they have been verified against the current MCP Server tool schema. Sponsored Brands product requirements depend on the selected campaign goal, ad format, and landing-page configuration. Sponsored Brands should not be described as universally requiring a minimum of three ASINs. Sponsored Display, now incorporated into Amazon's broader display ads offering, is a self-service Amazon Ads solution and does not inherently require Amazon DSP access. Sponsored Brands setup requirements depend on the selected campaign goal and ad format and may include brand, product, landing page, targeting, and creative inputs.

An auto campaign prompt instructs MCP to create a campaign where Amazon determines targeting: you supply the ASIN, budget, and bid. A manual keyword-targeted Sponsored Products campaign using exact match requires the keywords you want to target. Sponsored Products manual targeting can also use other keyword match types or product targeting. Auto campaigns are for discovery; exact match campaigns are for concentrating spend on known converters. Most sellers run both: auto to find new converting terms, and exact to exploit the ones already proven.

No. The review step is the control that separates AI-assisted campaign creation from unreviewed automation. For additional control, instruct the AI agent to present the proposed campaign structure and wait for explicit confirmation before requesting campaign creation. Verify that your MCP client and workflow support the intended approval behavior. Skipping the review risks campaigns launching with incorrect settings that waste budget from day one. For write-enabled campaign creation, asking the AI to show the proposed structure and wait for confirmation is a useful safety practice. However, users should verify that their AI client and MCP workflow actually enforce the requested confirmation behavior.