What 90 days actually buys you

For a clean B2B niche — meaning the prompt cluster is not yet contested by another active AEO programme — 90 days is enough to take a brand from 0% LLM visibility to 25–35%. The Gofaizen & Sherle case hit 33.3% on day 90 against six entrenched competitors, none of whom were running their own AEO programme.

If the niche is contested (a competitor has been on Scale or Enterprise for 6+ months already), 90 days gets you to 8–15% — meaningful but the catch-up case from the Why act now post.

This roadmap is the clean-niche version. For contested niches the structure is the same; the velocity is half.

Week 1: prompt research

Discovery week. Output: five seed prompts, ranked competitor list, frozen baseline.

  • Day 1–2: founder interview. Understand ICP, deal size, current pipeline sources.
  • Day 3–4: prompt mining (Searchable Agent, Profound, manual reverse-engineering). Surface 30–50 candidate prompts; rank by citation potential per hour-of-effort.
  • Day 5: pick five seed prompts. Run each in ChatGPT, Perplexity, Gemini, Claude and Google AIO. Screenshot the answer set for the frozen baseline.
  • Day 5: list the brands appearing in any answer. That is your competitor set.

By end of week 1 you have: five seed prompts, frozen baseline screenshots, named competitor list of 3–8 brands.

Week 2: competitor teardown

Output: four-signal score per priority page across each competitor.

  • Day 6–8: for each competitor, audit their top-15 cited pages. Score four signals: page length, schema completeness, named-expert byline, structural compliance.
  • Day 9: identify the gap. Most competitors land at 3-of-4 on schema, 2-of-4 on structure. Your target is 4-of-4 on all three.
  • Day 10: write the punch-list. 10–15 priority changes, BEFORE → AFTER format, ranked by citation potential per hour.

By end of week 2 you have: a competitor teardown spreadsheet and a punch-list ready to ship.

Weeks 3–4: schema and first rewrites

Output: schema deployed, first three priority pages restructured.

  • Week 3: deploy Article + FAQPage + BreadcrumbList + Person + Organization across the site. Validate every page against Schema.org and Google Rich Results Test.
  • Week 3: ship /llms.txt with brand summary, services, contact, out-of-scope sections. Set Cache-Control and Content-Type headers.
  • Week 4: rewrite the top three priority pages under the four-layer recipe. Hero, X-is-Y intro, Quick Facts (≥5 rows), H2-as-question with first-sentence direct answers, FAQ block (8–12 questions, ≤30 words each).
  • Week 4: deploy schema.org Person for every named expert. sameAs to LinkedIn + one verifiable other.

By end of week 4 you have: full schema stack live, three priority pages restructured.

Weeks 5–8: cluster build

Output: one pillar + six to eight detail articles, all under shared TZ.

  • Week 5: write the pillar (1,800–3,200 words, owns the highest-priority seed prompt).
  • Week 5–6: write four detail articles (800–1,500 words each). Bidirectional internal linking between pillar and details.
  • Week 7: write three more detail articles. Quality > volume; if a detail article is not earning its place in the cluster, drop it.
  • Week 7–8: shipping cadence stabilises at one to two articles per week.

Tracked prompts now expand from 5 seeds to 15–30 variants as the cluster surfaces them.

By end of week 8 you have: a complete topical cluster, internal linking validated, weekly tracking running.

Weeks 9–12: tracking and iteration

Output: first citations landed, pattern recognition, plan for month 4+.

  • Week 9: first AI Overview placements typically arrive (14–60 days after weekly 4 work). Document the first three.
  • Week 10–11: based on which content is citing and which is not, iterate. The 20% of pages getting 80% of citations get more depth; the rest get reviewed for structural compliance.
  • Week 12: 90-day report. Visibility, share of voice, citations per LLM, AOV of AI cohort if attribution is live. Plan for month 4–6.

By end of week 12 you should be at 25–35% visibility for a clean niche, 12–18% for a contested one.

Decision gates

Three gates the engagement passes (or fails) by:

  • Day 30: schema fully deployed, three priority pages restructured. If not, the engagement is behind and the next 60 days will not catch up. Pause and figure out why.
  • Day 60: first citations on at least one tracked prompt. If not, the structural compliance is incomplete (most common cause) or the niche is more contested than scoped.
  • Day 90: 25–35% visibility on tracked prompts (clean niche). If hitting 25%, continue. If below 15%, scope was wrong; renegotiate.

These gates are explicit at the start of every Answerly engagement. We refuse to run a programme without them.

What this maps to in our packages

The 90-day roadmap above is the Scale package executed cleanly: $4,800 / month base ($9,600 with the crypto/fintech multiplier). Three months total spend on the cycle: $14,400–$28,800.

If you want a slower, lighter version, Growth at $2,400 / month over six months covers the same ground at half the velocity. First citations land in months 3–4 instead of weeks 6–8, but the structure is identical.

The cheapest version: Starter audit for three months at $890 / month. You get the punch-list and the prompt research; your team owns implementation. If your team can ship under the playbook, this version produces the first citations within month two or three.