Automate Amazon Analysis Using Claude

List every Intentwise account I have access to. Show me account name, marketplace, and ad type in a simple table, and tell me which one had the most spend in the last 30 days.

Before we start: what kinds of questions can you answer using my Intentwise connection? Group your answer into (1) performance reporting, (2) AMC and audience analysis, (3) campaign structure, and (4) making changes to campaigns. One line per capability.

For [ACCOUNT], summarize the last 30 days against the previous 30 days: spend, sales, ACOS, ROAS, CPC, CTR, and conversion rate. Show each change in absolute terms and as a percentage, and flag anything that moved more than 15%.

ACOS moved in the last two weeks for [ACCOUNT]. Find the root cause: start at account level, then drill into campaign type, then campaign, then ad group, then target. Name the three specific entities that explain most of the change and tell me how much each one contributed.

Split the last 90 days of [ACCOUNT] into branded and non-branded search terms. Show spend, sales, ACOS, and share of total for each, broken out by month. Tell me the exact rule you used to classify a term as branded so I can correct it if it's wrong.

Take the last 90 days for [ACCOUNT] and build me a single-page dashboard: KPI tiles across the top for spend, sales, ACOS and ROAS; a weekly trend chart underneath; a top-10 campaigns table sorted by spend; and a section flagging every campaign running above a [TARGET ACOS]% ACOS. Dark theme. Publish it so I can share the link.

Good start. Now add a month-over-month comparison to each KPI tile, color ACOS red above [TARGET ACOS]% and green below it, and add a filter to switch between Sponsored Products, Sponsored Brands, and Display. Republish to the same link.

Write the Monday morning summary for [ACCOUNT] as if I'm sending it to a CMO with 90 seconds: what happened last week, the single biggest driver behind it, what we're doing about it, and the one number to watch this week. No jargon. Under 200 words.

Using AMC for [ACCOUNT], show me new-to-brand performance over the last 90 days: NTB orders, NTB sales, and NTB share of total, broken out by campaign type. Then tell me which campaign types are genuinely acquiring new customers versus harvesting demand we already had.

Using AMC, show me the most common ad exposure paths that led to a purchase for [ACCOUNT] in the last 60 days. Which combinations of ad types convert best, and what's the average number of touches before a purchase? Show me the query before you run it.

Look at the last 60 days of search terms for [ACCOUNT]. Find (a) terms with at least 2 conversions and ACOS below [TARGET ACOS]% that aren't already exact-match keywords, and (b) terms with more than [$ THRESHOLD] in spend and zero conversions. Give me two lists: harvest and negate. Preview only, do not make any changes.

Set up a scheduled task that runs every Monday at 8am: pull last week's performance for [ACCOUNT], compare it to the prior week, flag anything that moved more than 15%, and write it up as the CMO one-pager from P8. Send it to me when it's done.