How Search Worked Before AI

Before AI answers, search engines worked in three stages: crawlers discovered pages, an index stored what they found, and a ranking system returned a list of links for each query. Marketers competed for positions on that list, and success meant clicks. This lesson explains that model in plain language, because most of it still sits underneath AI search today.

In this lesson

• The three stages of search
• What the results page looked like and why position mattered
• How marketers optimised and measured
• What carries over into AI search
• Exercise and quiz

What are the three stages of search?

Google describes Search as working in three stages, noting that not every page makes it through each one (Google):

Google says there's no central registry of web pages, so it has to keep looking for new and updated ones, largely by following links from known pages (Google). That's why links, both from your own pages and from other sites, have always mattered for discovery.

What did the classic results page look like?

For years, a typical results page showed around ten organic links, often with ads above and below. Each link had a title, a URL and a short snippet. Over time, Google added more features: images, maps, videos, "People also ask" boxes and featured snippets, which quote a short answer from a page at the top (Google).

Position mattered because people tended to click the first few results. The page was a list; the user did the comparison.

What did marketers optimise?

Classic SEO had three broad pillars:

Google's own SEO Starter Guide still covers these basics: help Google find your content, organise it clearly, and write content people find useful. Keyword research, finding the words people type, guided what to write.

How did marketers measure success?

Tools like Google Search Console report impressions, clicks and average position (Google). The chain was simple: rank → click → visit → convert.

What changed, and what didn't?

Two things changed gradually, before generative AI arrived. First, more answers appeared directly on the results page, so some searches ended without a click. Second, people began asking longer, more conversational questions, partly because of voice assistants and mobile search.

Generative AI accelerated both. But the foundations didn't disappear. Google says its AI features use the same core systems as Search, and that pages need to be indexed to be eligible for them (Google). AI assistants that search the web also rely on crawlers. Lesson 2 explores what is new.

What carries over into AI search?

What changes most is the outcome: instead of a list the user compares, the AI often compares for them and names a few options.

A quick example

Illustrative. In 2015, someone searching "best running shoes for flat feet" saw ten links: review sites, retailers and a forum thread. They opened four tabs and compared. Today, the same person might ask an AI assistant and get a short answer naming three shoes, with sources. The review sites still matter, but now as sources the AI reads, not only as pages the person visits.

Why does this history matter for your work?

If you understand crawl, index and serve, you can diagnose most AI visibility problems. A brand missing from AI answers might have pages that can't be crawled, aren't indexed or simply aren't the best answer to the question. Those are the same three questions marketers have asked for decades.

Exercise

Map how 5 searches you made this week would have worked in 2015 versus today. For each, note: what you searched or asked, where you'd have clicked in 2015, what you actually did today (clicked a result, read an AI answer, asked a follow-up), and whether you visited any website at all.

Quiz

1. What are the three stages of Google Search?

a) Write, publish, promote b) Crawling, indexing, serving c) Bid, rank, click Answer: b

1. How does Google mainly discover new pages?

a) A central registry b) Following links from known pages c) Email submissions only Answer: b

1. What was the classic chain marketers measured?

a) Rank → click → visit → convert b) Post → like → share c) Ad → call → sale Answer: a

1. According to Google, what's needed to be eligible for AI features in Search?

a) A special AI file b) Being indexed and eligible for a snippet c) Paid placement Answer: b

1. What changes most with AI answers?

a) Crawlers stop existing b) The AI often compares options for the user c) Links no longer matter at all Answer: b

Next: Lesson 2: How AI changed discovery

Sources

• Google Search Central, In-depth guide to how Google Search works
• Google Search Central, SEO Starter Guide
• Google Search Central, Featured snippets
• Google Search Central, AI features and your website
• Search Console Help, Performance report