Problem
Authors and product sellers need one place to research demand, find opportunities, generate content, and run ads — without stitching Meta Ads Manager, Amazon, and SEO tools by hand.
Product roadmap
AI marketing intelligence for product research, SEO / AI Search content, and paid growth (Amazon-first, then Meta & TikTok). Born from the BookBooster discussion — validate on real Amazon books and Sponsored Products ROAS, with human approval before any ad changes.
Why this exists
Approach
Ship a lightweight Node.js web app first. Use Agoes’ Amazon books as the product set. Prioritise Amazon Ads (Direct Advertiser PoC) for clear spend→sale→ROAS; finish Meta in parallel; add TikTok after an account-region targeting spike. AI recommends — humans approve before publish.
Phases
MVP 1 — Marketing intelligence
Now / PoCCanonical product research loop. Validate with real Amazon book catalogue.
- Auth, workspace, product import (Amazon URL / website / manual)
- AI product intelligence, personas, keywords, competitors, search intent
- Opportunity scoring (“top things worth doing”)
- SEO briefs + AI article drafts + AI Search readiness
- Website SEO audit (crawl, findings, actionable opportunities)
- Dashboard · deployed PoC on Hetzner via ai.suherman.net
- Dashboard integration status (HaloRT + ad channels)
- Panduan PoC Amazon Ads (Bahasa) for Yamin & Om Agoes
- Bulk Amazon book import (CSV script)
- Polish UX with Agoes’ book catalogue end-to-end
- Export / share opportunity reports
- Decide commercial domain after validation
MVP 2 — Paid growth (Amazon-first)
NextConnect ad platforms into one place, analyse performance, and let AI recommend campaigns / targeting. Amazon before TikTok. Meta continues in parallel. Human approval before publish.
- 1 · Meta — Finish existing Meta connection (parallel)
- 2 · Amazon — Connect Agoes’ Amazon Ads account (Direct Advertiser)
- 3 · Amazon — Import books / ASINs and campaigns
- 4 · Amazon — Dashboard + keyword / search-term analysis
- 5 · Amazon — AI campaign recommendations
- 6 · Amazon — Campaign / bid / budget management (human approval)
- 7 · TikTok — Register developer / app access
- 8 · TikTok — Connect advertiser account (validate targeting region)
- 9 · TikTok — Campaign + creative analytics
- 10 · TikTok — AI targeting / creative recommendations
- 11 · Unified — Meta + Amazon + TikTok dashboard
- 12 · AI — Cross-channel Marketing Copilot
- Apply Amazon Ads API as Direct Advertiser (not Partner yet)
- Automatic → analyse → manual keyword campaigns loop for Sponsored Products
- Human approval before any budget / bid / launch write
- Spike AU TikTok Ads account: which US locations does that account expose?
MVP 3 — Attribution & AI optimisation
Later- Attribution across content + paid
- AI performance analysis and experiment suggestions
- Creative variants and budget recommendations
- Continuous optimisation loop (still approval-gated for live writes)
- Controlled auto-optimisation only after enough evidence
MVP 4 — Multi-channel scale
Later- Market research
- Google Ads, full TikTok / Amazon scale
- CMS publishing, Search Console, GA4 (OAuth + SEO audit enrichment)
- Social publishing helpers
- Optional Meta ads acquisition once product is solid (~budget experiment)
Near-term — This week
Cicil- Import 3–5 Amazon books into MarketPilot and re-run analysis
- Bulk import tooling (CSV script + roadmap auto-tick)
- Review opportunities + content drafts together
- Start Amazon Ads Direct Advertiser API application
- Lint in CI (ESLint alongside typecheck)
- Confirm Meta Business app request status
Near-term — Next 2–4 weeks
Cicil- Wire Amazon Ads OAuth + campaign / metrics sync (read path)
- Keyword / search-term dashboard for Sponsored Products
- AI recommendations + approve-before-apply flow
- Spike AU TikTok Ads account US targeting options