Amazon Ads MCP Server & AMC Expert Tool: What Agentic AI Means for Advertising Analytics

‍

Amazon advertising analytics is moving toward a very different operating model.

For years, getting a sophisticated answer from Amazon Ads often meant exporting reports, navigating API documentation, understanding Amazon Marketing Cloud schemas, writing SQL, validating queries, building dashboards and then translating the output into an advertising decision.

At unBoxed 2026, Amazon pushed directly at that complexity with two announcements: Amazon Ads MCP Server (Lite) and the AMC Expert Tool.

Amazon says MCP Server (Lite) gives AI agents a simpler way to discover and use Amazon Ads capabilities without loading the full advertising toolset upfront. The AMC Expert Tool goes deeper into analytics: it interprets a natural-language business question, reasons across Amazon Marketing Cloud schemas and advertising context, and generates an appropriate SQL query for the agent to use. Amazon Ads

The direction is significant:

BUSINESS QUESTION → AI AGENT → AMAZON ADS DATA → ANALYSIS → ACTION

Instead of advertisers adapting themselves to the technical structure of Amazon Ads, Amazon increasingly wants the technical infrastructure to adapt to the advertiser's question.

At Adorbix, we see the opportunity differently from simply saying, “AI can write SQL now.”

The real opportunity is:

less time retrieving data + less technical friction + faster analysis + more time deciding what to do with the answer.

What Is the Amazon Ads MCP Server?

MCP stands for Model Context Protocol, an open standard designed to let AI systems communicate with external tools and data sources.

Amazon introduced its Amazon Ads MCP Server in open beta earlier in 2026.

Amazon describes the system as a translation layer between an AI agent and Amazon Ads APIs. Instead of requiring an AI developer to manually coordinate many individual API requests, MCP tools can package common advertising processes into more complete workflows. Amazon Ads

For example, Amazon says MCP-based tools can support workflows involving:

campaign creation,

report generation,

account operations,

and international campaign expansion.

One Amazon example is creating an end-to-end Sponsored Products campaign through a single workflow rather than separately creating the campaign, ad group and ads through multiple API operations. Amazon Ads

The important idea is simple:

The MCP Server gives AI agents a structured way to work with Amazon Ads.

So What Is MCP Server “Lite”?

At unBoxed 2026, Amazon introduced a lighter architecture called Amazon Ads MCP Server (Lite).

According to Amazon, MCP Server Lite begins with only three core tools. Those tools allow the AI agent to discover which additional Amazon Ads capabilities it needs and then invoke them dynamically rather than loading the entire toolset into context from the beginning. Amazon Ads

That might sound like an engineering detail.

It has an important practical implication.

AI agents work better when they are not overwhelmed with hundreds of tool definitions they do not need for the current task.

Instead of telling an agent about every possible Amazon Ads operation before it answers a simple question, Lite can effectively allow the agent to ask:

What tools do I need for this job?

Then retrieve them.

That can make agentic advertising workflows:

simpler

more efficient

and potentially:

more reliable.

Why MCP Matters More Than Another Amazon Ads API

Amazon Ads APIs have existed for years.

The difference is the interface.

An API is built primarily for software.

MCP is designed around AI agents using software on behalf of people.

The old model might require:

Developer → API documentation → endpoint → authentication → query → output → application logic

The emerging agentic model looks more like:

Marketer → natural-language request → AI agent → MCP → Amazon Ads capabilities

That significantly lowers the technical barrier.

Amazon's broader Amazon Ads Agent product now explicitly says external AI agents and advertising platforms can connect to Amazon Ads through the MCP Server to create campaigns, run reports and execute multi-step workflows using prompts. Amazon Ads

For agencies, analytics teams and advertising technology companies, that's potentially a major infrastructure shift.

What Is the AMC Expert Tool?

The AMC Expert Tool addresses a different problem.

Amazon Marketing Cloud is enormously powerful, but getting value from AMC has traditionally required understanding its schemas and writing SQL.

That technical requirement creates a bottleneck.

A brand manager may know exactly what they want to understand:

Which customers saw video before clicking Sponsored Products?

But they may not know how to convert that question into a valid AMC query.

Amazon says the AMC Expert Tool can interpret a natural-language request, reason across AMC schemas and advertising context, then generate the correct SQL query so the broader AI agent can continue the workflow. Amazon Ads

That transforms AMC from:

“What SQL should we write?”

toward:

“What business question are we trying to answer?”

That is a much healthier analytics model.

Why AMC Has Historically Been Hard for Many Teams

Amazon Marketing Cloud lets advertisers perform deeper analysis than standard campaign reports.

It can help brands study questions involving:

customer journeys,

audience overlap,

frequency,

new-to-brand behavior,

cross-channel exposure,

conversion paths,

and sophisticated audience creation.

The problem is that Amazon's standard advertising managers, brand teams and even many PPC specialists are not SQL analysts.

So brands often end up with one of two situations.

Either AMC is underused.

Or every advanced question must wait for the analytics specialist.

The AMC Expert Tool can potentially reduce that dependency by translating business language into technically valid analytical work.

A Practical Example

Imagine an Adorbix client asks:

“Which customers were exposed to streaming video, then Sponsored Brands, and eventually purchased within 14 days?”

Traditionally, that might require an analyst to identify the relevant AMC tables, understand event structures, define attribution logic, write SQL, debug the query and interpret the output.

With an expert-tool workflow, the AI can help translate that question into AMC logic.

The human team can then spend more time deciding:

Does that path happen often?

Is it incremental?

What did the journey cost?

Should we invest more in streaming video?

That's where analytics actually becomes strategy.

Amazon Is Putting Nearly 1,000 Metrics Behind Natural Language

The AMC Expert Tool fits into a broader analytics shift announced at unBoxed.

Amazon says Amazon Ads Agent can expose nearly 1,000 advertising metrics through natural-language interaction.

Advertisers can ask about campaign drivers, new-to-brand performance, Long-Term Sales trends and other business questions, then receive visualizations and peer benchmarks without manually building every report. Amazon Ads

That changes the scarcity in Amazon advertising.

Previously, the scarce resource was often:

access to analysis.

Increasingly, the scarce resource becomes:

knowing which analysis is worth doing.

Agentic AI Does Not Make Analytics Strategy Less Important

This is the biggest misconception sellers and brands should avoid.

If anyone can generate a query by asking a question, it may seem like analytics expertise becomes less valuable.

In reality, the opposite can happen.

Consider these two questions.

Which campaign had the best ROAS?

versus:

Which campaign generated the highest incremental contribution from new-to-brand customers after accounting for repeat purchase and product margin?

AI can help answer both.

But only the second question reflects deeper business thinking.

The AMC Expert Tool improves execution.

It does not decide what the business should care about.

At Adorbix, this is exactly where human expertise becomes more important.

Bad Questions Can Still Produce Technically Correct Answers

Imagine asking:

Which campaign generated the most sales?

The AI may give you a perfectly correct answer.

But perhaps that campaign mostly captured customers who were already going to buy your branded product.

Another campaign generated less attributed revenue but introduced thousands of genuinely new customers.

The first answer is technically accurate.

The second campaign might still create more business value.

This is why agentic analytics should follow:

BUSINESS OBJECTIVE → QUESTION → QUERY → INTERPRETATION → ACTION

not simply:

QUESTION → NUMBER

The AMC Expert Tool Can Democratize Advanced Audience Building

AMC is not only an analytics environment.

Advertisers can use AMC insights to build more sophisticated audiences for activation through Amazon DSP.

Amazon recently published a case study involving Shokz and Amazon Ads partner Xnurta that shows how AI-assisted AMC SQL generation can work in practice.

Xnurta used an AMC SQL Generator expert tool through the Amazon Ads MCP Server to help translate natural-language business questions into validated AMC SQL. Amazon says this reduced some audience-analysis and audience-building workflows from days to hours and enabled less-technical team members to work more confidently with AMC. Amazon Ads

The resulting strategy helped the team develop audience segments based on signals such as purchaser lookalikes, category search behavior, PDP viewers, cart abandoners and lifestyle-related shopper intelligence. Amazon Ads

That is the more important story.

AI-generated SQL isn't valuable because SQL is annoying.

It is valuable because it allows more people to use advanced customer insight.

AMC Can Become a Faster Insight-to-Activation Loop

The traditional analytics model often looks like this:

Campaign runs

↓

Data exported

↓

Analyst investigates

↓

Insights presented

↓

Campaign team builds audience

↓

DSP campaign launched

That cycle may take days or weeks.

Agentic tools can compress it toward:

Business question

↓

AMC analysis

↓

Audience

↓

Amazon DSP activation

↓

Measurement

The faster the loop becomes, the faster brands can learn.

That becomes especially important during:

Prime events,

Black Friday,

product launches,

or fast-changing competitive periods.

This Is Where the MCP Server and AMC Expert Tool Work Together

The two tools solve different layers of the problem.

The MCP Server gives the AI agent access to Amazon Ads capabilities.

The AMC Expert Tool gives the agent specialized intelligence for complex AMC analytical tasks.

Think of it like this:

LayerRoleHumanDefines the business questionAI AgentCoordinates the workflowMCP ServerConnects the agent to Amazon Ads capabilitiesAMC Expert ToolTranslates advanced analytics questions into AMC logic/SQLAMCPerforms the analysisDSP / Amazon AdsActivates what the team learns

That's why Amazon described the announcements together at unBoxed. Amazon Ads

The goal is not isolated automation.

It is maintaining the workflow from question to action.

What Does This Mean for Agencies?

Agencies historically create value through a mixture of:

technical access,

platform expertise,

execution,

analysis,

and strategy.

Agentic AI changes that balance.

If campaign creation, report pulling and SQL generation become easier, agencies should spend less time selling:

“We know how to operate the platform.”

and more time delivering:

“We know how to create profitable strategy from the platform.”

At Adorbix, that means our role shifts even further toward:

business diagnosis

profitability analysis

cross-channel strategy

creative interpretation

customer acquisition

incrementality

rather than repetitive data extraction.

The Biggest Opportunity Is Faster Weekly Business Reviews

Imagine the traditional Monday PPC meeting.

Someone spends hours preparing:

Sponsored Products performance,

Sponsored Brands performance,

DSP,

new-to-brand,

placements,

search terms,

TACoS,

and creative performance.

The meeting starts after the analysis is built.

With agentic analytics, much of that preparation can become conversational.

A team could ask:

“What changed materially last week?”

Then:

“Which changes were caused by CPC, conversion, traffic or budget?”

Then:

“Which three issues have the biggest financial impact?”

The analyst can verify the output and focus the meeting on action.

That's substantially more valuable than simply producing dashboards faster.

But AI Still Needs Data Governance

Connecting external AI agents to Amazon Ads APIs creates obvious governance questions.

Who can connect an agent?

Which advertiser accounts can it access?

Can the agent only read data?

Can it create campaigns?

Can it change budgets?

Can it create audiences?

Amazon's MCP Server is currently available globally in open beta to Amazon Ads partners with active API credentials, according to Amazon's launch documentation. Amazon Ads

Brands and agencies should therefore treat agent access similarly to normal platform access.

Use:

least-privilege permissions

account-level controls

change logs

human approvals for high-impact actions

and:

regular audits.

Agentic does not mean uncontrolled.

Query Generation Still Requires Validation

Even if an AMC Expert Tool produces technically valid SQL, teams should still understand what the query is measuring.

Analytics errors are often not syntax errors.

They are definition errors.

For example:

What counts as new-to-brand?

Which conversion window is being used?

Is the analysis household-level or user-level?

Which exposure events qualify?

Are returned purchases included?

A query can run successfully and still answer a subtly different question than the marketer intended.

So Adorbix would maintain a review layer for important analyses.

AI removes syntax friction.

It should not remove analytical skepticism.

Don't Automate Metrics Without Profit Context

Amazon advertising systems can optimize beautifully around advertising metrics.

Your business economics may live elsewhere.

Suppose an AI identifies a campaign with excellent:

New-to-brand CPA.

But the product has:

8% contribution margin.

Another campaign has a higher CPA but sells a product with:

35% contribution margin.

Which deserves scale?

The answer isn't obvious from Amazon Ads metrics alone.

At Adorbix, we would combine agentic analytics with:

COGS

FBA

discounts

returns

contribution margin

customer lifetime value

before reallocating meaningful budget.

Natural-Language Analytics Could Change Who Uses AMC

Historically, AMC may have primarily been used by:

analysts,

large agencies,

enterprise advertisers,

and advanced DSP teams.

Natural-language query generation can expand access to:

PPC managers,

brand managers,

account strategists,

creative teams,

and potentially smaller advertisers.

That creates an interesting cultural change.

Advanced measurement no longer has to live inside one specialist department.

More people can explore the data directly.

The challenge becomes making sure everyone interprets it correctly.

Creative Teams Could Benefit Too

Analytics is not only for media buyers.

A creative strategist could ask:

“Which audience segments have high detail-page engagement but low conversion?”

Then investigate whether:

• creative is mismatched,
• product education is weak,
• the landing experience needs improvement,
• or the audience itself is wrong.

Another useful question might be:

“Which video-exposed audiences later convert through Sponsored Products?”

Now creative effectiveness can be analyzed within a broader customer journey instead of only through completion rate.

That's where AMC becomes powerful for CRO and creative strategy, not just attribution.

Agentic AI Could Help Catch Waste Earlier

Advertising waste often compounds because nobody notices quickly enough.

A campaign's CPC rises.

Conversion drops.

Spend continues.

By the time the weekly report is built, thousands of dollars may already be gone.

An AI agent connected through Amazon Ads infrastructure can theoretically help teams identify exceptions much faster.

For example:

“Alert me when spend rises 25% while attributed conversion declines.”

Or:

“Flag campaigns whose new-to-brand CPA exceeds our target for three consecutive days.”

The value is not simply faster reporting.

It is earlier intervention.

The Adorbix Agentic Analytics Framework

At Adorbix, we'd implement these tools around five stages.

ASK

Start with a business problem—not a metric.

Examples:

“Why is customer acquisition getting more expensive?”

or:

“Which campaigns are generating incremental demand?”

ANALYZE

Use Ads Agent, AMC and the Expert Tool to investigate the question.

VALIDATE

Check:

definitions,

query logic,

time periods,

and business assumptions.

CONNECT

Add external economics:

COGS,

fulfillment,

returns,

margin,

and customer value.

ACT

Only then:

increase budget,

change audiences,

adjust bids,

or restructure campaigns.

This prevents AI from making technically impressive but commercially weak decisions.

How Adorbix Can Use the Amazon Ads MCP Server

For brands and partners, MCP opens up several potential workflows.

Adorbix could use agentic systems to accelerate:

automated weekly campaign reviews

cross-account anomaly detection

campaign setup preparation

international campaign expansion

audience research

performance reporting

AMC analysis

and:

optimization recommendations.

The important word is:

accelerate.

We would not replace the strategy layer.

We would remove repetitive execution underneath it.

How Adorbix Can Use the AMC Expert Tool

The AMC Expert Tool can make sophisticated analytics more accessible across an account team.

Instead of every business question waiting for a SQL specialist, strategists can explore questions such as:

Which audience paths create the highest new-to-brand conversion?

How many exposures typically happen before purchase?

Does streaming video contribute to later Sponsored Products sales?

Which audiences are overexposed?

Which customer segments respond best to remarketing?

The Expert Tool handles more of the technical translation.

Adorbix focuses on:

what the answer means.

The Questions Brands Should Start Asking

As analytics becomes easier, brands should raise the quality of their questions.

Instead of:

What was our ROAS?

Ask:

Which campaigns generated sales we were unlikely to receive without advertising?

Instead of:

Which audience had the best conversion rate?

Ask:

Which audience created the best contribution after media cost?

Instead of:

How many people saw the video?

Ask:

Did video exposure change later purchase behavior?

Better tools should lead to better questions.

Otherwise brands will simply generate old reports faster.

A Practical 30-Day Agentic Analytics Pilot

During the first week, identify five recurring Amazon Ads analyses that consume the most team time.

For example:

weekly performance summaries,

new-to-brand analysis,

campaign anomaly identification,

audience research,

and DSP reporting.

During week two, test natural-language workflows and compare the outputs with your current manual process.

During week three, introduce AMC questions that previously required analyst support.

Validate every query and document the definitions used.

During week four, measure:

time saved

accuracy

issues discovered earlier

and:

business decisions improved.

If AI only produces more reports, the pilot has failed.

If it reduces manual work and improves decision speed, expand it.

Agentic Analytics Readiness Checklist

Before scaling an MCP or AMC Expert Tool workflow, confirm that the organization has clear business KPIs, Amazon Ads API permissions, AMC access where needed, documented metric definitions, an approval process for campaign changes, profitability data outside Amazon and human ownership of final strategic decisions.

The technology can become sophisticated very quickly.

The operating governance should become sophisticated with it.

Frequently Asked Questions

What is the Amazon Ads MCP Server?

Amazon Ads MCP Server is Amazon's implementation of the Model Context Protocol for advertising. It connects compatible AI agents to Amazon Ads API capabilities so agents can create campaigns, run reports and execute multi-step advertising workflows through natural-language prompts. Amazon Ads

What is MCP Server Lite?

Amazon announced MCP Server (Lite) at unBoxed 2026. It begins with three core tools that allow AI agents to discover and invoke additional Amazon Ads capabilities only when needed rather than loading the entire toolset upfront. Amazon Ads

What is the AMC Expert Tool?

Amazon says the AMC Expert Tool interprets natural-language advertising questions, reasons across Amazon Marketing Cloud schemas and advertising context, and generates an appropriate SQL query for the requesting agent. Amazon Ads

Do marketers still need SQL?

The technical requirement is becoming significantly lower. The AMC Expert Tool and related expert tooling can generate AMC SQL from natural language, although teams should still validate important queries and understand what the resulting analysis actually measures. Amazon Ads

Can MCP create Amazon advertising campaigns?

Yes. Amazon says its MCP Server supports tools for workflows including campaign creation, reporting and expansion. One example can build an end-to-end Sponsored Products campaign by coordinating several underlying operations. Amazon Ads

Who can use the Amazon Ads MCP Server?

Amazon's February 2026 documentation says it is available globally in open beta to Amazon Ads partners with active API credentials. Amazon Ads

How does this connect with Amazon Ads Agent?

Amazon Ads Agent is Amazon's unified AI-powered advertising platform. Amazon says external AI agents and platforms can connect to its advertising capabilities through the Amazon Ads MCP Server. Amazon Ads

Can AMC insights be activated in DSP?

Yes. AMC analysis can support audience creation and DSP activation. Amazon's Shokz/Xnurta case study shows AMC-derived segments being activated through Amazon DSP and optimized through Performance+. Amazon Ads

Does agentic AI replace advertising analysts?

It can automate substantial technical and repetitive work, but the business still needs humans to define objectives, validate analysis, connect advertising to profitability and decide which actions make strategic sense.

Can Adorbix help brands use these tools?

Yes. Adorbix can combine Amazon Ads MCP workflows, AMC analysis, DSP, Sponsored Ads and external profitability data so agentic AI accelerates decision-making without allowing automation to replace strategic judgment.

Key Takeaways

Amazon's unBoxed 2026 analytics announcements move Amazon Ads further into the agentic AI era.

The Amazon Ads MCP Server provides infrastructure that lets external AI agents interact with Amazon Ads capabilities. Amazon Ads

The newer MCP Server Lite simplifies agent access by beginning with three core tools that dynamically discover additional capabilities as needed. Amazon Ads

The AMC Expert Tool tackles one of Amazon Marketing Cloud's biggest barriers by translating natural-language business questions into appropriate AMC SQL. Amazon Ads

Meanwhile, Amazon Ads Agent now exposes nearly 1,000 metrics through natural-language analytics, reinforcing the broader move away from manual reporting toward conversational analysis. Amazon Ads

The technical barrier is falling.

That makes:

question quality,

interpretation,

governance,

and:

profitability context

more important than ever.

Final Takeaway: AI Is Making Amazon Ads Analytics Easier—So Strategy Becomes the Scarce Skill

The old advertising advantage was partly technical.

Know the report.

Know the API.

Know the AMC schema.

Know SQL.

Know how to extract the answer.

Agentic AI increasingly handles those mechanics.

The future workflow looks more like:

ASK → ANALYZE → VALIDATE → ACT → MEASURE

That does not eliminate advertising expertise.

It changes where expertise lives.

At Adorbix, we believe the advantage shifts toward teams that know:

which business question matters,

which metric actually answers it,

whether the analysis is economically meaningful,

and what action should follow.

Because when every advertiser can generate a sophisticated AMC query in seconds, the competitive advantage will not be:

“We know how to write the query.”

It will be:

“We know what to ask—and what to do with the answer.”

Amazon's MCP Server and AMC Expert Tool can make the analytics engine dramatically faster.

Adorbix's job is to make sure that faster engine is pointed toward profitable growth.

‍