How One Manager Operates 775 ASINs Across 9 Brands
Human judgment × AI cognition × Amazon Ads API execution — results from Month 1 of Pulson Agents.
AT A GLANCE
- Customer
- KISS Products · Multi-brand beauty (9 sub-brands, 775 ASINs)
- Marketplace
- Amazon US
- Amazon Ads Products
- Amazon Ads API (Campaigns, Reporting), Sponsored Products, Sponsored Brands
- Solution
- Pulson Agents
- Period
- Feb 2026 – Mar 2026
HEADLINE RESULTS
Challenge
KISS is a multi-brand beauty advertiser operating 9 sub-brands, 775 ASINs, and 67 product groups. At takeover, every campaign was consolidated into just 9 broad-category portfolios — making it impossible for humans to see which product groups were over-invested and which were missing opportunities. Monthly ad spend swung by nearly 4× across seasonal peaks and valleys. This was a scale problem manual operations could not solve.
Solution — Human in the Loop
The innovation isn't automation. It's the deliberate division of labor between people, AI, and APIs:
- Strategy — The PulseAd PSM decides why. - Intelligence — AI on top of a domain ontology detects patterns and diagnoses. - Execution — The Amazon Ads API turns decisions into real campaign changes in real time.
Because the system runs on an explicit domain ontology — "product group," "campaign purpose," "investment matrix" — AI interprets data as operational context, not raw numbers. Pulson Agents restructured 9 portfolios into 67 product groups × purpose-based segments and now executes 36,644 automated bid adjustments per month across 3,429 targets.
Results
First month live: ROAS +41%, CPC −27%, Ad Sales +53%. Year-over-year: ad spend −32%, tACoS −1.5pp. An estimated 15–25 hours per week of manual work disappeared. The same architecture runs from a 50–200 ASIN single brand to a 775 ASIN multi-brand portfolio — only the brand parameters change.
The Takeaway
Scale problems are solved by better operating systems, not more headcount. AI doesn't replace human judgment — it extends its reach and speed.
