You’ve got the framework: Discoverability, Clarity, Authority, Trust — Found → Understood → Cited → Recommended. You’ve run the audit and you know your weakest layer. Now you turn it into a plan.
This is the part most teams never reach. They read about AI search, nod, and go back to their Google dashboards — because measurement is the single biggest gap1 in GEO, and a strategy you can’t measure is a strategy you can’t defend at budget time. So this plan is built around two things: a sequenced 90 days of work, and a scoreboard you can put in front of leadership.
Principles before the plan
Work the layers in order. Don’t chase recommendations while crawlers are blocked. The plan front-loads the foundation (Discoverability, Clarity) because the later layers don’t pay off without it.
Pick a focus, not a to-do list of forty items. Your audit told you the weakest layer. Over-invest there. A plan that tries to fix everything fixes nothing.
Measure from day zero. Your baseline audit is day zero. Re-run the same prompt set every month so every change is measured against a real “before.”
The 90-day plan
Three phases, roughly 30 days each. Adjust the emphasis to wherever your audit said you’re weakest — but keep the sequence. Copy it into your planner or download the CSV to start filling in owners and dates.
| Phase | Days | Focus (layers) | The goal | Key moves |
|---|---|---|---|---|
| 1 — Foundation | 1–30 | Discoverability + Clarity | Be found and understood | Baseline audit; unblock AI crawlers; fix server-rendered facts; publish llms.txt; consistency pass on core facts; Organization/Product schema + sameAs; pursue Wikidata; correct the worst stale claims |
| 2 — Authority | 31–60 | Authority | Get cited | Build citation target list from the audit; earn brand mentions (digital PR); get into “best/alternatives” content; genuine community + YouTube presence; close the biggest citation-share gaps vs competitors |
| 3 — Trust + cadence | 61–90 | Trust + reporting | Get recommended, and make it repeatable | Build review presence on key platforms; respond to reviews; set up sentiment monitoring + alerts; stand up the monthly report; re-run the full audit and compare to day zero |
A few notes on doing this realistically:
- Phase 1 is mostly technical and editorial — much of it is same-week work (crawlability, schema, a facts consistency pass) that unlocks everything after it. Don’t skip it because it’s unglamorous; it’s the highest ROI 30 days you’ll spend.
- Phase 2 is the slow one. Authority is earned over months, so 30 days is about starting the engine — pitching, publishing, participating — not finishing. Expect movement to show up in months two and three.
- Phase 3 turns the project into a practice. The deliverable isn’t just better trust signals; it’s a standing monthly report and an owner, so AI Alignment becomes a habit instead of a one-off sprint.
What to track: the AI Alignment KPIs
Old SEO KPIs (rankings, sessions, backlinks) don’t capture this. Track metrics that map to the layers. Keep the set small enough that you’ll actually maintain it.
| KPI | What it measures | Layer | How to read it |
|---|---|---|---|
| AI Visibility Score (mention rate) | % of your prompt set where you’re mentioned | Discoverability | Your headline “are we present” number |
| Share of answer | Your mentions vs competitors’ on the same prompts | Authority | Competitive presence — the new “share of voice” |
| Citation share | % of cited sources in your category’s answers that are about/from you | Authority | Whether you’re in the sources deciding the category |
| Accuracy rate | % of answers about you that are factually correct | Clarity | Rising = your entity work is landing |
| Sentiment | Positive / neutral / negative tone of mentions | Trust | The recommendation proxy |
| Recommendation rate | % of decision prompts (“best X”) where you’re actively recommended | Trust | The bottom line — chosen, not just named |
Two rules for these to stay honest: always measure them against the same prompt set month over month, and always show them next to competitors. A 40% mention rate means nothing alone; “40% vs the category leader’s 75%” is a story leadership understands instantly.
Reporting: make the invisible visible
The reason to track all this is to show it. A simple monthly one-pager beats a sprawling dashboard nobody opens. Put on it:
- The six KPIs, this month vs last vs the day-zero baseline.
- You vs your top 2–3 competitors on share of answer and recommendation rate.
- Three real answer screenshots — the most important buying questions, showing whether you’re present, accurate, and recommended.
- What we changed, and what moved — tie the work to the metric.
- Next month’s focus — the one layer/gap you’re attacking.
The screenshots do the heavy lifting. A chart proves the trend; a screenshot of the model recommending a competitor makes it visceral. If you want to go beyond a one-pager, we’ve written a full guide to building a GEO reporting dashboard, and there’s a worked benchmark example of what tracked data looks like over a quarter.
Common pitfalls
- Tracking too many models or prompts to sustain. Three models, 20–30 prompts, run consistently, beats a giant set you abandon in month two.
- One-and-done. AI answers drift constantly. A single audit is a snapshot; the value is in the trend line.
- Optimising vanity prompts. Winning on “what does [your brand] do” feels good and means little. Win on the decision prompts where buyers choose.
- Going wide instead of deep. Picking at all four layers at once. Fix the foundation, then climb.
- Shortcuts on Authority and Trust. Manufactured mentions and fake reviews are detectable and can convert an authority gap into a trust crisis.
Where to go from here
That’s the series. You have the framework, the audit, a tactic set for each layer, and a 90-day plan with KPIs. If you want the wider tactical library that sits alongside this course, our complete GEO playbook goes deeper on individual plays. And the whole thing starts and ends in the same place: run the audit, pick your weakest layer, and give it 90 honest days.
Take the whole course with you — every worksheet, plan, and checklist in one file:
Download the AI Alignment ToolkitSources
-
Search Engine Land — 8 GEO metrics to track in 2026 ↩
-
Search Engine Land — How to plan for GEO in 2026 and evolve your search strategy ↩
-
Ahrefs — How to track AI Overviews: mentions, citations, and click loss ↩