SEO in the AI-search era

The honest picture

Search still sends more visitors to most content sites than any other channel, and it sends fewer than it did. Both are true. One widely cited study put zero-click searches at about 68% of US Google queries in early 2026, and click-through rates fall sharply when an AI Overview appears above the results. If your plan was to write a thousand short answers to simple questions and collect the traffic, that plan is over. Google now answers those itself.

What is not over: people searching for things that need judgement, comparison, experience, or a tool. Those queries still click through, because a summary cannot do the job. So the strategy is to become the page a summary points to, and the page a person opens when the summary is not enough.

What still works, and works better than before

Topic hubs, not scattered posts

A hub is one strong page on a subject with a set of linked articles beneath it: the hub explains the whole topic, each spoke goes deep on one part, and every spoke links back. Search engines read the structure as expertise. Readers read it as a place to stay. This whole Journey section is a hub: an index page and ten steps that link forward and back. Build every topic that way, and stop publishing orphan posts that link to nothing.

E-E-A-T, which is a fancy name for proof

Google's quality guidelines ask for Experience, Expertise, Authoritativeness and Trust. In practice, for a solo site, that means:

• A real author with a real name, a photo, and an about page that says what they have actually done.
• First-hand evidence in the content: your screenshots, your numbers, your mistakes. This is the part AI cannot fake and the part readers now look for.
• Sources linked, dates shown, prices verified and dated.
• Contact details, a privacy page, and a clear line about how the site makes money.

None of this is a ranking trick. It is what makes a page worth citing, which is now the same thing as ranking.

Schema

Structured data tells machines what a page is. Add Article, Person (for authors), Organization, FAQ where there genuinely are questions, and Product or Review schema where you review things. It takes an hour to set up once with any of the site options from step three. AI systems that summarise the web use the same signals; a clearly typed page is easier to quote correctly.

Speed and no clutter

A page that loads in under two seconds on a phone, with no pop-ups covering the content, still wins ties. It also earns better ad rates later, because ad networks measure the same thing. Test in Google's PageSpeed tool and fix the images first; they are almost always the problem.

What changed

• AI Overviews and AI Mode answer informational queries in place. Do not compete with the answer box on "what is X" questions. Write for "which X for my situation", "X versus Y", and "how I did X", where the reader needs more than a paragraph.
• Being cited is the new ranking. ChatGPT, Perplexity and Google's AI features quote sources. Pages get cited when they state a clear claim, back it with a specific number or example, and are easy to attribute (author, date, organisation). Vague pages get summarised without credit.
• Bing matters again. ChatGPT's web search draws heavily on Bing's index alongside its own. Verify your site in Bing Webmaster Tools, submit your sitemap, and turn on IndexNow. It takes twenty minutes and most of your competitors have not done it.
• Long-tail volume is smaller. Keyword tools show falling numbers on simple queries because people now ask assistants instead. Do not panic about a chart; look at whether the queries that still send you clicks are growing.

What to stop doing

• Publishing thin "what is" pages.
• Stuffing a keyword into every heading. Machines read meaning now.
• Buying links. It has not worked for a decade and it still gets sites removed.
• Rewriting the same article every quarter to "refresh" it without adding anything. Update when there is something new, and say what changed.

Tools: which ones earn their price

Google Search Console is free and is the only tool that shows you what Google actually does with your site: queries, clicks, positions, indexing errors. Set it up on day one and look at it weekly. It keeps sixteen months of history, so export quarterly if you want more.

Semrush and Ahrefs do the same core jobs: keyword research, competitor gaps, backlink data, site audits. Semrush Pro is currently $139.95 a month, Ahrefs Lite $129, both with annual discounts. Neither is worth paying for until you have twenty articles published and a question the free tiers cannot answer. Both free tiers are usable for a few queries a day. When you do pay, pay for one month, do a full month of research, export everything, and cancel. The data does not go stale in a quarter.

Surfer analyses top-ranking pages for a query and scores your draft against them. It is useful for making sure a comparison article covers what readers expect and harmful if you let it turn writing into box-ticking. Entry plans currently run around $79 to $99 a month depending on billing, and the plan names changed this year, so check the current page. Skip it until you are publishing several long articles a week.

The tool nobody lists: your own site search logs. What people search for on your site, and do not find, is your next ten articles.

The weekly habit

• Monday: open Search Console. Note queries where you rank in positions 5 to 15. Those pages are one improvement from page one; improve them before writing anything new.
• Check one competitor's newest pages and ask what question they answered that you have not.
• Add internal links from your three most-visited pages to your newest one.

Fifteen minutes. Most sites never do it.

Checklist

• Search Console and Bing Webmaster Tools verified with a sitemap submitted
• Content organised into at least one hub with linked spokes
• Author page, sources and dates on every article
• Article and Person schema added and validated
• Positions 5 to 15 queries reviewed and one page improved this week