Checklist or SOP

How to Choose an Amazon Ads MCP: 5 Questions to Ask

TL;DR

To choose an Amazon Ads MCP, ask five questions: who can do what, whether an engine keeps optimizing between chats, whether bulk changes are previewed, how many tokens each answer uses, and how long it takes.

  • Ask about permissions first: read and write granted separately, per person and per ad account, with expiry and revocation.
  • Decide whether you want your AI to make every bid decision or to steer an engine that keeps working between chats.
  • For sales events, require hourly data, bulk edits, a server-enforced old to new preview and a change log.
  • Ask how much raw data each tool returns. Finished results mean fewer tokens, so ask for measured figures and repeat the test yourself.
  • Start read-only, then add write access once the provider documents its write controls.

Get every answer in writing, then test it on one real ad account. The same questions work for Amazon's own Ads MCP server, ads platform MCPs and self-hosted open-source servers. New to the term? Start with what an Amazon Ads MCP is.

Why do agencies and brands need different answers?

Agencies and brands ask the same five questions but need different answers. An agency runs many client accounts with a mixed-seniority team; a brand runs fewer accounts, with more people outside the ads team asking for answers.

QuestionWhat an agency needsWhat a brand needs
PermissionsEach client walled off; juniors propose, seniors applyFinance and leadership read-only; the ads lead can change
Engine behind itClient goals run between check-ins, so one manager covers more accountsTargets held all day, not only when someone asks
Bulk changesOn event days, hourly reads and budget and bid changes across clientsFix a budget running out at 2 p.m. on Prime Day in one step
Token useAccount managers on standard AI plans don't run out mid-weekBig catalogs stay inside the AI's working memory
Response timeA Monday review across a dozen accounts in one sittingAnswers fast enough to use in a meeting

1. How precisely can you control who does what?

Permission control limits how much damage a mistake or a forgotten connection can do. For an agency, a connection that sees every client is a liability; for a brand, finance should never be one prompt away from changing a bid.

AreaWhat to askStrong answerWeak answer
Sign-inDoes each person sign in with their own login?Per-person sign-inOne shared login or access key
SeparationAre read, edit, create and delete granted separately?Each is its own permissionAll-or-nothing access
ScopeCan a connection be limited to selected ad accounts?Per connectionPer login only
CeilingCan a connection exceed the person's own role?NeverUnclear
LifetimeDoes it expire, and who can revoke it?Expiry you choose; revoke any timeLong-lived, no expiry

Anthropic advises reviewing the permissions each custom connector requests. A typical split: analysts and leadership read-only, account managers with campaign edits on their own clients, and one senior lead who can create campaigns. The full order is in read-only vs read-write MCP.

2. Is there an optimization engine behind the MCP?

An engine is optional, but it changes how the MCP works. An MCP with one keeps working toward your goals between conversations; an MCP without one acts only when someone asks. If every bid and budget move has to come from a prompt, your team still does the optimization, by typing instead of clicking.

Direct control is enough for small accounts, custom agents and teams that want to approve every bid. Engine-backed control suits large catalogs and agencies covering many clients.

  • Ask: Does an algorithm keep optimizing between chats? Can I set the goal, target ACOS, budget and bid limits for a group of campaigns from the assistant, and choose which levers it may use? Can I see what it changed?
  • Strong answer: goals and guardrails set per group of campaigns from the chat, a continuous engine within them, and its changes in a log the assistant can read.
  • Weak answer: autonomous optimization that turns out to be a scheduled prompt starting from scratch each run.

Set target ACOS from each product's break-even ACOS (margin before ad spend), not a round number. A $40 product with $26 in cost and fees has a $14 margin, so break-even ACOS is 35%; a target of 25% to 28% leaves profit after ads. See AI-managed group levers and personality for the controls.

3. Can the MCP act on live data in bulk, hour by hour, and safely?

Bulk write on live data is where an Amazon Ads MCP pays for itself, especially on Prime Day, Prime Big Deal Days or Black Friday, when a hero campaign runs out of budget at 2 p.m. and category CPCs jump in the evening. An agency running a dozen clients can't catch that through exports and console clicks.

Event-day momentWhat you askWhat you apply in one bulk change
Budgets running out earlyWhich campaigns spent over [80]% of today's budget by [2 p.m.] with ACOS under target?Raise those budgets [30]%, previewed old to new
Deal ASINs converting above planWhich keywords on deal ASINs converted above [2]x their normal rate in the last [3] settled hours?Raise their bids [15]%, capped at a maximum you set
Spend without salesWhich campaigns spent over $[200] since midnight with no orders and 20+ clicks?Lower bids, previewed first; pause only the clear outliers
Event endsList every budget and bid raised today.Set them back in one change
  • Ask: Which data is hourly, and how recent? Which edits work by ad type? How many changes fit in one request? Is campaign creation all-or-nothing? Is the old to new preview enforced by the server? Is the result read back and logged by operator?
  • Strong answer: a written table by ad type, stated bulk limits, a preview the server enforces for budgets, bids and archiving, a read-back and a searchable log.
  • Weak answer: read-only, so every recommendation ends in console work; or writes where the model asks before it acts depends on the prompt.

Any change you confirm goes live and spends real budget. Archiving on Amazon can't be undone, so pause instead. For the working flow, see bulk edit Amazon campaigns with AI.

4. How much of your AI plan does each answer use?

Token use sets how many questions your team gets from the Claude or ChatGPT plans it already pays for, and the main driver is what the MCP hands back. An MCP that passes Amazon's raw reports straight to the model makes it pull reports, page through rows and calculate ACOS itself. Anthropic's engineers note that with direct tool calls, every intermediate result must pass through the model, and large results can overflow its working memory on big accounts.

  • Ask: For last 30 days ACOS by campaign with the biggest movers, how many tokens does a typical answer use? Do your tools return raw report rows or finished results?
  • Strong answer: compact, pre-calculated results, with token counts measured on real questions and a published test method.
  • Weak answer: answers built from hundreds of raw rows, and long reports cut short.

MCP compatibility rarely separates providers anymore; token use does. For an agency, it is the difference between finishing the week's client questions on a standard plan and hitting the limit by Wednesday.

5. How long does each answer take?

Response time decides whether an Amazon Ads MCP fits the working day. A question that takes 15 minutes won't be asked in a client call. Most latency comes from round trips: report requests, waiting for reports, paging and calculation inside the chat.

  • Ask: How long does a typical question take from prompt to answer? How many tool calls does it make? What happens with long date ranges or several accounts?
  • Strong answer: measured timings on real questions, and few tool calls per answer.
  • Weak answer: real time, with no figures.

How do the three kinds of MCP score on the five questions?

Amazon's own server is stronger for billing and first-party reach, a platform MCP for an engine to steer, and self-hosted for full control with a developer.

QuestionAmazon's own Ads MCP serverAds platform MCPSelf-hosted open-source
1. PermissionsThrough API credentials and partner setupUsually tied to platform logins and rolesWhatever you build
2. Engine behind itNo group-level engine listed in Amazon's announcementSometimes, through the platform's automationNo
3. Bulk write safelyWorkflow tools; created campaigns need reviewVaries; check that the server enforces previewsWhatever you build
4. Token useDepends on how much raw data each tool returnsVaries; ask for measured figuresUsually raw API output
5. Response timeDepends on report round trips and pagingVaries; ask for measured timingsDepends on your build
Stronger whenYou need first-party reach, billing or account settingsYou want an engine to steerYou want full control and have a developer

Many teams run both. See Amazon's own Ads MCP server vs a platform MCP.

How do you run the evaluation on one account?

Run it on one real ad account with a read-only connection, before anything is signed or any write access is granted.

  1. Send the five questions in writing, and ask for documentation of permission and write controls as well as a demo.
  2. Repeat the token and speed test yourself: connect read-only, ask the same three recurring questions through each MCP in fresh chats with the same model, and record minutes and usage. Check every number against the console.
  3. Test one write on a paused campaign: preview, confirm, check the read-back, then find the change in the log.

How does Xnurta MCP answer the five questions?

Xnurta MCP is an ads platform MCP built on Xnurta's optimization engine. Its answers are below, limits included, so you can score it the same way as any other option.

QuestionXnurta MCP
1. PermissionsRead and write permissions granted separately; all or selected ad accounts; never more than the person's own Xnurta permissions; expiry after 30, 90 or 180 days or 1 year; revocation takes effect immediately
2. Engine behind itAI-managed groups: Xnurta's AI optimizes each group toward a goal and target ACOS with only the levers you turn on, plus a bid range and date-range schedules for events
3. Bulk writeOld to new preview for budgets, bids and archiving, enforced by the server rather than the AI's prompt, then a read-back; operation logs by operator; live-campaign edits vary by ad type, with the most options for Sponsored Products; Sponsored Products campaigns created in bulk, all-or-nothing, optionally paused
4. Token useDesigned to return compact, finished results instead of raw report rows, which keeps cost and token use down; measure it on your own account with step 2 above
5. SpeedDesigned for fewer round trips, because the analysis happens behind the server; time it with step 2 above
Where it's weakerNo billing or account-settings tools, which Amazon's own server lists

To stop a change, decline the preview, pause the AI-managed group or revoke the connection in the Xnurta platform. Generally available from 27 Oct for Sponsored Ads. Write access is turned on per account by your CSM. Learn more on the Xnurta MCP page, or book a demo to run these five questions live.

Next question: Is it safe to connect Amazon Ads to ChatGPT or Claude?

Video transcript

Sources & references

FAQ

Frequently asked questions

How long should an Amazon Ads MCP trial last?

Plan on four weeks. Spend week one read-only, checking numbers against the console, then add campaign edits in week two and budgets in week three. Test automation or campaign creation in week four, and avoid starting a trial in the two weeks before Prime Day or Black Friday.

What does an Amazon Ads MCP cost to run?

An Amazon Ads MCP has three cost parts: AI plan seats for each user, the tokens each answer uses, and server access. Token use is the part you can influence most, because an MCP that returns finished results instead of raw report rows uses fewer tokens per answer. Ask every provider whether MCP access is priced separately.

Does an Amazon Ads MCP cover Sponsored Brands and Sponsored Display?

Coverage varies by server and by action. Most servers read all three ad types, but edits and campaign creation often have the most options for Sponsored Products. Ask each provider for a written table of reads, edits and creation by ad type.

What if an Amazon Ads MCP change goes wrong?

Most changes can be reversed by editing back, and the change log shows what changed, when and by whom. Archiving on Amazon can't be undone, so pause instead and keep delete permissions off. Revoke the connection right away if a change looks wrong, then review every change it made that day.

Can one MCP connection serve several clients?

It can, but it shouldn't. Give each client its own connection, limited to that client's ad accounts, with its own permissions and expiry. Then a mistake or a departing employee affects one client only, and each client's change log stays clean for client reviews.

Does an Amazon Ads MCP train AI models on my client data?

Training depends on your AI plan, not the MCP. Anthropic and OpenAI say their business and API plans aren't trained on by default, while consumer plans differ. Use a business plan for client work and ask the MCP provider for its written data policy.

Xnurta MCP

Ask your Amazon Ads account a question today.

Xnurta MCP is for Xnurta customers only. Create your MCP connection directly in the Xnurta platform. New to Xnurta? Book a demo and we'll walk through your account with you.