The first 90 days of AI search for a stretched-thin founder

If you're a founder reading this with 200 tabs open and three people on the marketing team, you do not need a 6-month AI search transformation programme. You need a 90-day plan that fits inside what your team can actually do this quarter. Here it is.

The premise: 30 days of measurement, 30 days of structural fixes, 30 days of content the AI engines can quote. Each phase is ~10 hours of work per week from one person, not a transformation programme.

Days 1–30, Measure before you move

The most common waste of the first month is "rushing to fix things." You don't know what's broken yet. Spend the first 30 days establishing a baseline.

Week 1, Pick the 20 queries that matter

Sit down with whoever knows your buyer best and write the 20 queries your highest-intent customers actually type or speak. Not "best AI tools." Specific things. "AI search consultant for Indonesian SMEs." "How to optimise a CPG site for AI Overviews." Twenty rows, named buyers behind each one.

These are your tracked set for the next 90 days. Resist adding more, focused beats broad here.

Week 2, Check where you stand across four surfaces

For each of those 20 queries, check:

• Google classic search, what's your position?
• Google AI Overviews, does an AI Overview appear at all? If yes, are you cited?
• ChatGPT, when you ask the query directly, are you mentioned in the answer? With what description?
• Perplexity, same question

A spreadsheet works. Don't pay for a tool yet, at 20 queries the manual check takes a couple of hours and tells you more than a dashboard ever will.

Week 3, Audit how you're described

Now ask each AI engine directly: "What does [your company] do?" "Best [your category]?" "Who should I hire for [your service]?"

Write down what they say. Look for:

• Wrong descriptions (more common than you'd think)
• Inconsistent descriptions across engines (one says you do X, another says Y)
• No description at all (you're not in their graph yet)

This tells you whether you have a visibility problem or an understanding problem. They're different fixes.

Week 4, Pick the one number to track

From everything above, pick one metric that becomes your North Star for the next 60 days. Most teams should pick "% of the 20 tracked queries where we appear in an AI engine's answer."

It's slow to move. It's imperfect. But it forces the right behaviour, make us quotable, make us correctly described, make us cited, instead of the chase-rankings reflex.

Days 31–60, Fix the structure

Now you have a baseline. The next 30 days fix the structural things that decide whether you're findable at all.

Week 5, Person and Organization schema

Add Person and Organization JSON-LD to every page on your site. Include sameAs links to your verified LinkedIn, Crunchbase, and any other profile where you're recognised. This is one afternoon of work and it materially feeds the AI engines' entity graph.

If you don't know how, ask your developer. If you don't have a developer, every modern CMS has a plugin. The hard part isn't the implementation, it's deciding what your canonical company description is across all those surfaces, and committing to it.

Week 6, Direct answers in the first 100 words

Walk through your top 10 commercial pages and rewrite the opening so the answer to the page's title is in the first 100 words. No throat-clearing intro. No "in today's fast-moving landscape." A direct, quotable definition or answer.

This is the single highest-leverage change you can make. AI engines lift the opening. If your opening is fluff, you don't get cited.

Week 7, Fix the obvious technical breakage

Run your top 30 pages through a free crawler (Screaming Frog free tier, or Sitebulb's free version). Fix:

• Missing or duplicate title tags
• Missing or duplicate meta descriptions
• Broken internal links
• Missing alt text on meaningful images
• Pages that should noindex but don't

This is plumbing. None of it is interesting. All of it matters.

Week 8, Consolidate or split cannibalising pages

If two pages on your site compete for the same query, AI engines hedge, and Google often picks neither. Decide which page should win. Either consolidate (merge into one stronger page and redirect the loser) or differentiate (rewrite one to target a related but distinct query).

Days 61–90, Make yourself quotable

The final 30 days, you start producing content the AI engines can actually quote.

Weeks 9–10, Two pillar answers

Write two definitive pieces on the topics most central to your category. Long-form, opinionated, with clear definitions and an obvious point of view. Each piece should answer the most-asked question in its sub-area.

Forget "SEO content", write the piece that, if you sent it to a smart buyer, would make them think "OK, these people get it."

Weeks 11–12, Three "decision" posts

Write three shorter posts in decision-framework format: "When to do X vs Y," "Should you do X or Y, a founder's framework," "Three mistakes I see in X." AI engines disproportionately cite this format because it gives them clean, structured chunks to lift.

(This post is one of those, for the record.)

What you'll have at day 90

A measured baseline. Person and Organization schema across the site. Top 10 pages with direct answers in the first 100 words. The obvious technical breakage fixed. Two pillar pieces. Three decision posts.

Re-run the week 2 measurement. The numbers should move. If they haven't, something's wrong with the diagnosis, not the work, and that's worth a call.

If you've read this and want a second brain on the diagnosis before you spend 90 days, book a free 30-minute call. I'll tell you whether this plan fits your situation or whether you need a different shape entirely.