Is blogging still worth it when AI answers everything?

Short answer: yes, but not the kind of blogging most businesses do. Broad explainer posts — "what is content marketing", "a beginner's guide to email" — are finished. An AI summary answers those at the top of the page and nobody clicks through. Narrow posts that answer one specific question a buyer asks just before they spend money are worth more than they were two years ago.

Here is what changed, and what to do about it.

What actually changed

Two things, and they pull in opposite directions.

The first is that people stopped clicking. Roughly 68% of Google searches in the United States now end without anyone visiting a website, up from about 60% in 2024. When Google shows an AI summary at the top, click-through to the results below drops by around 60%. If your traffic fell off a cliff and you could not work out what you did wrong, this is probably it. You did not do anything wrong.

The second is less obvious and more useful. AI assistants do not only quote the pages that rank first. When someone asks a complicated question, the system breaks it into smaller questions and goes looking for an answer to each one. Recent analysis found that only 38% of the pages cited in Google's AI summaries were ranking in the top ten — down from 76% a year earlier. Almost a third of citations came from pages ranking beyond position 100.

Read that again, because it is the whole opportunity. You no longer have to outrank anybody to get quoted. You have to be the clearest answer to one narrow question.

What to stop writing

• Anything beginning "the ultimate guide to".
• Definitions of terms your customers already understand.
• Round-ups of tips that exist on forty other sites.
• News commentary, unless you have something first-hand to add.

These lose to AI summaries every time, because a summary does exactly that job and does it faster.

What to write instead

Write the things a person types when they have nearly decided to buy and are looking for a reason not to. In practice that means:

• Comparisons. "X versus Y for a small business." Buyers search these constantly and they are hard for an AI to answer without citing somebody.
• Prices and what drives them. Almost nobody publishes real numbers, which is exactly why the pages that do get quoted.
• What goes wrong. "Why does my X keep doing Y." First-hand experience is the one thing a summary cannot manufacture.
• Worked examples. Not a blank template — a filled-in one, with the real numbers, even if the client name is removed.

How to write it so it gets quoted

Research on this is thin but consistent on a few points. Answer the question in the first two or three sentences under each heading, before you explain anything. Include real numbers and name your sources. Use quotations from named people where you have them. Keep each section able to stand alone, because a machine may lift one section and nothing else.

Stuffing keywords does nothing. It measurably underperformed doing nothing at all in the one controlled study I have seen on this.

The uncomfortable part

Your traffic numbers will probably keep falling even if you do all of this correctly. That is the shape of the whole web right now, not a verdict on your writing. Stop measuring visits and start measuring inquiries. Ten visitors who arrive already convinced are worth more than a thousand who bounce.