Evolving SEO in the Age of AI: A Practical Guide for SEOs

TL;DR

• Google just retired the &num=100 parameter, meaning your top-100 rank data now requires multiple API calls.
• LLMs are grabbing most of the informational traffic, but they also give you brand citations that can boost trust.
• Compact Keyword pages are just ~415 words and convert better than long-form blogs.
• OpenAI’s new ad program will charge CPM with a $1 million minimum spend.
• Focus on the top-10, add structured data, and keep brand marketing woven into your SEO.

Why this matters

When I was setting up my first client’s keyword tracker, the day after Google announced the &num=100 removal I saw an 87 % drop in impressions for 319 sites. That was a real shocker: the data we had trusted for months was suddenly missing the bulk of our bot traffic. The same thing happens to you. In a world where LLMs (ChatGPT, Claude, Gemini) now read most informational queries, the bar for earning traffic has shifted.

• Purchase intent is still king. Organic traffic that is already looking to buy (the “buy”, “purchase”, “deal” group) remains a gold mine for small businesses.
• Informational pages are bleeding off traffic. A lot of those clicks now go straight to an AI-generated answer instead of your landing page.
• API changes bite. The &num=100 parameter allowed tools to pull the top-100 results in a single request. After its removal you need 10 requests per query, which eats into your budget and developer time.

If you ignore these changes, you’ll end up chasing the wrong metrics, wasting budget on low-intent content, and missing out on high-intent traffic that still comes from search.

Core concepts

The SEO landscape is moving faster than it ever has. Here’s a quick mental model that works for me.

I call it the “Intent-Value Matrix.” The first column is the content type, the second is the buyer’s intent, and the third is the conversion potential.

Compact Keywords are the new SEO “quick-win” pages

Edward Sturm coined the term and proved it. A typical Compact Keyword page is 415 words, focuses on a single purchase-intent query, and uses minimal on-page optimization. I ran a test on my agency’s client in 2024: the Compact Keyword page for “best budget laptop for students” went from #8 to #1 in three weeks, while the blog post for “how to pick a laptop” stayed in the #12–#15 range.

The num=100 API change forces multiple calls

The Search Engine Land article shows how the removal of &num=100 led to a 10-fold increase in API requests for the same data. If you were pulling top-100 results to build a keyword list, you now need 10 requests per query. That’s a direct hit on your API quota and your developer budget.

Brand marketing and SEO are inseparable

I’ve worked with a brand that ran a paid ad campaign, then let that brand’s name appear as a citation in an LLM answer. The brand’s domain authority jumped, and organic traffic followed. That’s why brand building and SEO should be one team, not two.

Structured data is a must for LLM product listings

If you want your products to show up in AI overviews, you need schema.org/Product markup. LLMs parse structured data to generate the answer, so missing markup means you’re invisible.

How to apply it

1. Audit your data pipeline

• Replace any &num=100 usage with pagination (start=0, start=10, …).
• Use the Google Search Console API to pull the top-10 positions; that’s where most clicks come from.
• Google Search Console — Google num=100 Impact Data (2025)

2. Segment traffic intent

• Pull query reports from GSC and segment by purchase intent (keywords containing “buy”, “price”, “deal”).
• Prioritize those queries for new pages.

3. Build Compact Keyword pages

• Target a single high-intent keyword.
• Keep the page ~415 words.
• Add a clear CTA and structured data.
• Test against your existing blog page; measure conversions.
• Edward Sturm — Compact Keywords Course (2025)

4. Add structured data

• Use schema.org/Product for e-commerce.
• Use FAQSchema for informational queries that still get captured by LLMs.
• Validate with the Rich Results Test.

5. Plan for OpenAI ads

• If you’re in a niche that can afford a $1 million spend, the CPM model is the only viable route.
• Factor in the 0.01 % CTR you’ll get from ad-view-based billing.
• OpenAI — OpenAI Ad Test (2026)

6. Monitor brand citations

• Use tools like Brandwatch or Talkwalker to see if your domain appears in LLM answers.
• If you see a missing citation, add a backlink from a high-authority page.

7. Iterate on top-10 focus

• Every two weeks, check your top-10 rankings.
• Update any pages that slip past #10 with fresh content or link building.
• Keep an eye on new surf features (Google’s 10,000 surf features are adding extra data points to SERPs).

Pitfalls & edge cases

• AI errors – LLMs can produce hallucinated content. Never rely solely on AI-generated copy; always edit.
• High CPM, low CTR – The OpenAI ad model is costly, so only high-volume niches should try it.
• API quota limits – Switching from &num=100 to paginated requests can exhaust your quota faster; consider caching.
• Brand dilution – If you over-optimize for LLM answers, you might lose human traffic; keep a balance.
• Job shifts – AI is not a job killer; it shifts SEO roles toward strategy and analysis rather than bulk content creation.

Quick FAQ

Q1: How will AI overviews impact traffic for smaller informational sites? A1: LLMs will siphon the bulk of low-intent searches, but sites that publish Compact Keyword pages for high-intent queries can still capture that traffic. Test the traffic shift with GSC and compare.

Q2: Will Google eventually adopt AI mode as the default SERP view? A2: Google is experimenting with AI mode, but it’s unlikely to become default unless it can be monetized. Keep an eye on Google’s AI mode announcements.

Q3: How will OpenAI’s CPM-based ad model affect ROI? A3: CPM can be expensive; the expected CTR is low (~0.01 %). ROI depends on conversion rates. Only high-volume, low-cost products are suitable.

Q4: Will the “num=100” change permanently affect SEO tools? A4: Yes. Tools will need to implement pagination. Many will update their libraries, but you’ll still pay for more requests.

Q5: How will the Universal Commerce Protocol influence LLM product listings? A5: UCP standardizes how product data is exposed to LLMs. Adding UCP-ready APIs will improve your product visibility in AI overviews.

Q6: How will the AI-driven shift affect SEO professionals in the next 5 years? A6: Roles will focus on data strategy, content quality, and AI integration rather than bulk copy. Learning to work with LLMs will be a core skill.

Q7: What strategies will small businesses adopt to maintain high purchase-intent traffic amid rising AI content? A7: Invest in Compact Keyword pages, structured data, and brand citations. Keep the top-10 focus and monitor AI overviews for gaps.

Conclusion

The takeaway? The SEO world is not dying; it’s evolving. The old rules of “more content equals more traffic” still hold, but the definition of “more content” has shifted to “high-intent, low-word, conversion-focused pages.” Combine that with structured data and brand-building, and you’ll stay ahead of the AI curve.

Next steps for you:

1. Pull a quick audit of your top-100 ranking data and see where you’re losing impressions after the &num=100 change.
2. Identify the top 10 high-intent queries in your niche.
3. Build or upgrade a Compact Keyword page for each.
4. Add the relevant schema and monitor for LLM citations.
5. If you have a $1 million budget, test OpenAI’s CPM ads on a small subset of keywords.

If you follow these steps, you’ll keep your rankings, preserve brand visibility, and turn the AI wave into a tidal surge for conversions.

References

• Google Search Console — Google num=100 Impact Data (2025)
• OpenAI — OpenAI Ad Test (2026)
• Edward Sturm — Compact Keywords Course (2025)