Welcome back to The PPC Den! This episode we're pulling back the curtain on how a real Amazon agency runs behind the scenes from AI to accountability.
Michael sits down with Joshua Rawe of AmpliSell to talk about how they personally use AI (spoiler: Claude Opus is the daily driver), why "average" AI output is a trap if you don't already have great ideas, and how they built an ongoing, AI-powered account audit so nothing falls through the cracks between reviews.
We also get into the messy reality of client goal-setting: how to pin clients down on revenue vs. margin vs. ACOS before you ever touch a bid and the #1 mistake killing Amazon brands in 2026: not killing dead products and not launching new ones fast enough. Plus, why winning on Amazon today means winning off Amazon first.
00:00 - Intro
02:30 – Personal vs. offline AI use, LM Studio for local LLMs
04:45 – Staying on top of Amazon changes with Claude Opus 4.8
07:00 – Amazon's title character limit change (75 characters) as a case study
08:15 – How information flows from AI digest → Slack → SOP → knowledge base
11:00 – Trust, delegation, and not micromanaging the last mile
13:30 – The "wasted spend check" problem: audits that never get repeated
15:45 – Building an ongoing, AI-run account audit ("audit-proof accounts")
18:00 – Client retention goals and the emotional cost of getting fired
19:30 – Weekly OKR feedback loops with leadership
21:15 – Nailing down client goals: revenue vs. margin vs. ACOS
24:00 – Biggest mistake in 2026: not killing bad products / not launching new ones fast enough
26:30 – Anker example category expansion vs. one hero product
28:00 – AI and copywriting/images: client pushback then client demand
29:30 – Highest-leverage activity: advertising + off-channel marketing
31:00 – Why Amazon is a demand-capture engine, not demand-creation
