Amazon Ads analytics · AI co-pilot · MCP

Every Amazon ad dollar, explained.
Every next move, proposed.

Adsinq is the analytics workspace for Amazon advertisers. Campaigns, search terms, hourly spend, Brand Analytics, inventory and profit sit in one place, and an AI layer reads all of it to propose bids, negatives and budgets. Nothing changes until a person approves it.

Book a demo See pricing
app.adsinq.com / halvora-home-us / home
Adsinq Home: performance by brand, headline tiles for sales, spend, ACOS and TACOS, and every other metric one click away

Screens are from the product with a fictional brand ("Halvora Home"), and every name and number is illustrative.

30+analysis views, from campaign to search query
Hourlyspend and sales via Amazon Marketing Stream
SP · SB · SDplus Brand Analytics, SP-API and Walmart
100%of AI changes previewed, bounded and undoable

01 · The product

One workspace for everything Amazon tells you, and what it doesn't.

Built by operators who managed hundreds of Amazon accounts and got tired of eight dashboards that never agreed. Every view reconciles to the same numbers and uses the same filters, date range and comparison period.

app.adsinq.com / halvora-home-us / home
Overview: performance by brand and headline tiles Campaigns: daily trend chart and a campaign table with budgets and budget used Search terms: which queries are targeted or negated, with match type and campaign N-gram analysis: single words ranked by spend with ACOS against target Hourly performance: today's spend per hour against a normal day, and a 30-day heatmap Market share and funnel: weekly search volume with impression, click, cart-add and purchase share To do: budget-pacing proposals with the reason for each change, waiting for approval

02 · What's inside

Four sections, no missing pieces.

Live in production Beta in the beta app Soon on the roadmap

Advertising

  • Portfolios, campaigns, ad groups Live
  • Targeting: keywords, products, auto Live
  • Placements and placement allocation Live
  • Audiences and AMC Beta
  • Hourly (AMS) and dayparting Beta
  • Creatives performance Live
  • Campaign builder Soon

Search and market

  • Search terms and shopper intent Live
  • Search Query Performance (SQP) Live
  • N-gram analysis, 1 to 6 words Live
  • Market share and funnel Live
  • Competitors: market watch, head-to-head Live
  • Rufus and Alexa prompts Beta
  • Keyword rank tracker Soon

Inventory and profit

  • Products: performance, cross-sale, intelligence Live
  • Inventory and restock signals Live
  • Profit with COGS and fee overrides Live
  • Returns by reason and ASIN Live
  • Subscribe & Save Live
  • Walmart channel Beta

Automate

  • To do: AI and rule proposals Beta
  • Rules and automations Beta
  • Exclusions and negation lists Beta
  • Skills: reusable AI playbooks Beta
  • Change logs with undo Live
  • Multi-account, teams, targets Live

03 · AI

Bring your own model. Keep your hand on the wheel.

Adsinq exposes your Amazon data and a small set of safe actions to AI through the Model Context Protocol (MCP). Connect any MCP-capable AI assistant and it can read campaigns, search terms and SQP, then draft changes. Every change comes back as a preview first.

Your AI assistantAdsinq MCP connected
Which search terms wasted the most in Halvora Home over the last 14 days?
get_ad_performancelevel: "search_term", days: 14, orders: 0, sort: "spend"
Three terms spent $187 with no orders: "herb saver" $84, "salt pig" $61 and "tea towels" $42. None of them matches what you sell in those campaigns. Shall I negate them as exact?
Yes, do it.
add_negative_keywordspreview: will_set 3 · skipped 0 · limited_by_bounds 0
Here's the preview: 3 negative exact keywords across 2 campaigns. Approve to apply?
Approve.
add_negative_keywordsapply: true · 3 applied · logged to Change logs
Done. All three are live and can be undone from Change logs at any time.

Guardrails, built in

  1. Preview first. Every write tool returns what it would do: what's set, what's skipped, what's limited.
  2. Safety bounds. Account-level bid ceiling, maximum raise per change and budget step. The AI can't go past them.
  3. A person approves. Changes apply only after an explicit yes, in chat or in the To do queue.
  4. One-click undo. Bid and budget changes reverse from Change logs, with who, what and when.
  5. Editor role only. Read-only teammates can ask questions. Only editors can change anything.

Tools your AI can call

account_overviewget_ad_performancelist_campaignslist_keywordsquery_datasetget_change_historyset_bidsset_budgetspause_entitiesadd_keywordsadd_negative_keywordsundo_change

read   write, preview first

Where AI works in Adsinq

Each process reads the same reconciled data you see on screen and lands in the To do queue as a proposal with its reasoning.

Beta

Daily recommendations

Overnight, rules and AI check every account and turn what changed into proposals in To do, each with its reason and evidence.

Beta

Search-term harvest & negation

Converting queries get promoted to exact targets. Ones that spend without orders get negated, with the evidence attached.

Beta

Bid suggestions

Revenue per click against your target ACOS gives the bid each target has earned, within your safety bounds.

Soon

AI bidding

Continuous bid management that learns from your approvals, moving from proposing to acting inside limits you set.

Beta

Dayparting

Hourly AMS data shows when shoppers convert. Bids follow the hours that pay and pull back from the ones that don't.

Beta

Budget pacing & cap rescue

Campaigns that run dry while under target ACOS get more budget. Over-target ones stay capped, and each gets a reason.

Live

Shopper-intent clustering

Search terms are grouped into intents like own brand, competitor, use case and feature, so you can see which intents convert and which leak.

Soon

Listing & Rufus insights

Real shopper questions from Rufus and SQP vocabulary, turned into title and bullet suggestions you can check against the data.

04 · Pricing

Two plans. Tokens included.

Dashboards and data are unlimited on both plans. Tokens only count when AI does the work: answering questions, writing recommendations, drafting changes.

Analytics

For brands and teams who run their own ads

$199/ month

Up to 3 Amazon accounts · +$49 per extra account

3MAI tokens / month included
top up at $15 per 1M
  • Every analytics view: advertising, search, market, inventory, profit
  • Hourly data via Amazon Marketing Stream
  • Brand Analytics: SQP, market share, competitors
  • AI assistant on your data, with recommendations in To do
  • MCP connector for your own AI assistant
  • Unlimited seats, roles and saved views
  • CSV and PNG export on every table and chart
Start with Analytics
Most hands-off

Analytics + Management

We run your Amazon Ads with Adsinq and AI

from$1,500/ month

or 4% of monthly ad spend, whichever is higher

15MAI tokens / month included
top up at $15 per 1M
  • Everything in Analytics, for all your marketplaces
  • AI bid management and rules, run by us within your bounds
  • Search-term harvesting, negation and dayparting
  • Budget pacing and cap rescue, checked every day
  • A senior strategist on your account, never a junior
  • Weekly report and monthly review, plus a shared Slack channel
  • Full change log, so you see every move we make
Talk to us about Management

What's a token?

A token is a small piece of text the AI reads or writes, roughly ¾ of a word. A typical question about one account, with its answer, uses 20–60k tokens. A full overnight review of one account uses about 150k.

What uses them

The built-in assistant, overnight recommendations, intent clustering and drafted changes. Browsing dashboards, filtering, exporting and data refreshes never use tokens.

Your own model

Connect your own AI through MCP and the model's usage goes on your AI provider's plan, not on your Adsinq tokens. Your data still goes through the same guardrails.

Which Amazon data do you connect to?

Amazon Ads API (Sponsored Products, Brands and Display), Amazon Marketing Stream for hourly data, Selling Partner API for orders, inventory, fees and returns, and Brand Analytics for Search Query Performance. Walmart Connect is in beta.

How fresh is the data?

Daily reports load overnight and are typically one day behind, the same as Amazon's console. Hourly spend, clicks and sales arrive through Amazon Marketing Stream during the day.

Can the AI change my campaigns without asking?

No. Every change is a preview until someone with the Editor role approves it. Account safety bounds cap bids and budget steps, and bid and budget changes can be undone from Change logs.

What happens when I run out of tokens?

Analytics keeps working. AI features pause until the next month or until you add a top-up ($15 per 1M tokens). We email you at 80% so it never comes as a surprise.

Is there a contract?

Analytics is month to month. Management starts with a 3-month term so there's time to restructure and measure, then continues month to month.

05 · Who builds it

Operator-built, not pitch-deck-built.

Adsinq is led by a founder with close to a decade of hands-on work in Amazon advertising, across hundreds of brand accounts in more than twenty categories. That work covers Sponsored Products, Brands and Display, Amazon Marketing Cloud, Brand Analytics and the retail side behind the ads: inventory, margins and listings.

The product started as the internal tooling behind that work: bid models, data pipelines and the reports used to make real budget decisions. We still manage accounts with it every day, so every view has to earn its place by helping someone decide what to do next.

9+years in Amazon PPC
300+accounts worked on
24categories

See your own account in Adsinq.

A 30-minute walkthrough on your data: where the money goes, what the AI would change first, and what it would cost.

Book a demo

hello@adsinq.com