Multi-account content & outreach orchestration for a creator network

§01Challenge

The client manages a network of 47 brands around the creator economy (lifestyle, fitness, tech). Each brand has accounts on four platforms: Instagram, TikTok, X and YouTube. Prior structure: three operations people, each manually running ~16 accounts. At that headcount they reached ~120 publications per week. Goal: 1,200/week while maintaining quality and 0 bans.

Standard tools (Buffer, Hootsuite, Later) broke at this account count, mixed accounts between marketers, struggled with TikTok and offered no engagement loops. Most of the market stops here. The client did not want to stop.

§02Approach

Operator console in Next.js with three modules: content calendar (overview of 47 brands × 4 platforms in one grid), publishing engine (official API where it exists, safe scraping for TikTok and certain Instagram features), engagement queue (DMs, comments, mentions scanned by an intent classifier; ambiguous ones go to the operator queue).

Each brand has a "personality profile" — activity hours, frequency, tone, content categories. This is not human simulation — it is the operational fact that brands are run by different team members with different schedules.

Compliance layer: every action auditable, daily "what the system did" report per brand, kill switch per brand per platform (enforced with one click). Isolated session architecture means a ban on one brand does not threaten the other 46.

§03Outcome

Output 11.6× higher with the same operational headcount. Per-publication time dropped from ~45 minutes to <3 minutes (review and approve). 60–70% of DMs handled automatically, the rest queued.

Zero bans in 6 months of production. One shadow-ban on TikTok (the client pushed more aggressively than we recommended), recovered with a one-week pause and pattern adjustment.

The client closed a new $1.4M/yr advertising partner mandate on the back of the first months of data from the system.

Stack

Metrics

• 47 — Brands / accounts
• 11.6× — Output lift
• <3 min — Per-publication time
• 60–70% — DMs auto-handled
• 0 — Bans
• 6 mo — Autonomous runtime