Personal brand: 90 days

A three-month program: define who the person is for whom, set a cadence they can keep, ship it week by week through ProPost, and review with real numbers at day 30, 60 and 90. Consistency and focus beat volume.

Before you start

1. get_account: plan, time zone, drafts left per month, weekly cap, approval setting, autonomy. The plan must fit the cadence (drafts, cap).
2. list_channels: pick 1 or 2 primary channels (usually LinkedIn, plus one of X, Instagram or Threads). More than 2 spreads thin.
3. list_personas, get_posting_schedule, list_ideas: what exists. If voice or schedule is missing, run propost-brand-voice or propost-account-setup-audit first.
4. get_channel_stats per primary channel with range: "90d": the baseline (followers, impressions, engagement rate). Record it. A new account may return little; say so.
5. list_posts (published, limit 50): what has worked, from engagement.

Workflow

1. Positioning interview (one message): what they want to be known for; the audience; the 90-day goal in one outcome (inbound leads, a job, a speaking slot, 1,000 followers who matter); what they can prove; time per week they can give.
2. Write the positioning statement and show it: "I help [specific audience] [achieve specific outcome] through [method or angle], based on [proof]." Example: "I help seed-stage founders price B2B software without discounting, based on 60 pricing projects." Proof must be theirs. Iterate once.
3. Set the pillars (3 to 4) with propost-content-pillars logic and seed the idea bank with add_idea.
4. Set the cadence to what they can sustain, based on slots and cap:
   Light: 3 LinkedIn posts a week.
   Standard: 3 LinkedIn + 3 short posts on a second network.
   Heavy: 5 + 5; only if they really have the time and ideas. Posting times: suggest_posting_times; set_posting_schedule (schedule level, replaces the week, confirm first).
5. Show the 90-day map (table below) and the AI draft budget: roughly 1 generate_draft per post you do not write yourself. At 3 posts a week for 12 weeks that is up to 36; compare with drafts left per month and plan to write some with create_draft or alternate.
6. Execute week by week. At the start of each week (or all four weeks of the month at once if they prefer batching), follow propost-weekly-content-plan: plan table, yes, drafts, check_post, add_to_queue. Don't pre-write 12 weeks; plans change with what works.
7. Monthly review at day 30, 60, 90 (below).
8. Close at day 90 with a summary against the baseline and a next-quarter plan.

90-day map

Monthly review

1. get_channel_stats for each primary channel, range: "30d"; compare with the previous-period totals returned. Report: followers gained, impressions, engagement rate, and the change.
2. list_posts (published, limit 100, from set to the month start): rank by engagement (reactions, comments, impressions, link clicks, saves). Name the top 3 and bottom 3 posts and what they had in common: pillar, format, hook, day and time.
3. Say what the data cannot show: if engagement is null for many posts, the numbers are incomplete; do not infer. Never invent numbers.
4. Recommend 3 changes: what to do more of, less of, and test next. Keep one variable per test (hook style, format, posting time).
5. Update the plan: reweight pillars, adjust slots with set_posting_schedule (with confirmation), bank new ideas with add_idea.

Example finding: "Story posts averaged 3x the comments of how-to posts; both Tuesday posts beat Friday ones. Next month: 2 stories, 1 how-to a week, move Friday to Thursday 08:30."

Writing guidance

• Say the positioning aloud in the first line of the profile and in recurring posts, but not as an ad. Show it with proof.
• Vary the formats inside one pillar; repetition reads as a template.
• Alternate give and ask: roughly 4 posts that teach or tell for each post that invites (a call, a download, a conversation).
• Replies and comments on others' posts do as much for visibility as posting; you cannot do this through ProPost, so ask the person to spend 10 minutes a day on it.
• Small numbers are fine. Report 40 new followers who match the audience rather than inflating.

Guardrails

• Targets are the person's, not promises. Don't guarantee follower growth or leads.
• No invented credentials, clients or results in the positioning or posts.
• Don't schedule or change the posting schedule without a yes.
• Never publish_now as part of the plan.
• Don't compare to other people's accounts using made-up benchmarks; use their own baseline.
• Stay within AI drafts and the weekly cap; state them each month.

If something fails

• quota_exceeded: write posts with create_draft; recheck the reset date.
• weekly_limit: reduce the cadence or leave extras as drafts.
• no_slots / queue_paused: set_posting_schedule (schedule level, with confirmation) or schedule_post with exact times.
• Empty stats: the channel may be new or not synced; say so and rely on post counts and the person's own observations.
• autonomy_limit: everything stays a draft; list what is waiting.
• Anything else: see ../CONVENTIONS.md.