Which Blog Topics Are Becoming Obsolete Because of AI-Generated Summaries?

Which topics are losing their value?

What these share is that the reader wanted a piece of information, not an experience of the page. When the information is common and short, the answer takes the click.

Which topics hold up?

• Original data: surveys, benchmarks, price indexes, analyses of your own anonymised records. Only you have them, which gives answers a reason to cite you.
• First-hand testing and experience: "we ran 12 CRMs through the same onboarding and timed each step". Experience is hard to fake and hard to summarise without credit.
• Tools: calculators, configurators, checkers and templates, which require a visit.
• Detailed comparisons with checkable facts, including prices, limits and where each option fits.
• Complex how-tos: multi-stage processes with screenshots, edge cases and troubleshooting, where readers need to follow along.
• Local and current specifics: prices in a named city, regulations in a specific state, what changed this month.
• Judgement and opinion from real experts, clearly argued, especially on contested questions.
• Case studies and proof, which recommendation answers need.

These are the pages AI answers cite and people still click, because they offer something the summary points to rather than replaces. Which formats are cited most often is covered in which content formats get cited most often by AI.

Are basic “What is X?” articles still useful for AI search visibility?

Yes, in a narrower role. They are no longer a good traffic play for common terms, but they still do three jobs:

• Defining your own terms and category. If you coined a method, a product category or a named framework, your page is the natural reference for it. Answers need somewhere to learn what it means. How named frameworks become citation handles is covered in how naming an original framework can create a citation handle for your brand.
• Defining terms where your expertise adds something. "What is a chart of accounts?" is generic. "What is a chart of accounts for a dental practice, and which accounts do practices usually need?" adds specifics an answer can cite.
• Anchoring a topic on your site. A short, accurate definition page that links to your deeper guides helps readers and crawlers understand how your content fits together. A glossary can do the same job, as covered in whether glossaries help AI understand a website's topics and terminology.

What does not work any more is a 2,000-word "What is X?" article padded to rank, with the definition buried under an introduction. If you keep these pages, put the definition in the first sentence, add what only you can add, and keep them short.

How do you sort your existing blog?

1. Export every post with its traffic over the last 12 months, compared with the year before.
2. Tag each one by type: definition, fact, tips, short how-to, news, roundup, comparison, data, experience, tool, case study, opinion.
3. Check what AI answers do with the main question each post targets. Does an AI Overview answer it completely? Does ChatGPT cite anyone?
4. Decide for each post: - Keep posts with original data, experience, tools or proof. - Upgrade posts on good topics that lack something only you can add. Add your data, examples, screenshots or a comparison table. - Merge near-duplicates into one stronger page, with redirects. - Retire posts that are fully replaceable and bring no traffic or links, redirecting where there is a close match.
5. Re-plan new content around the topics that hold up.

How to update pages that already rank so assistants cite them is covered in how to rework top-ranking articles so AI assistants cite them.

Should you stop writing about common topics altogether?

No. Common topics are where buyers start, and being cited on them still builds recognition, even without clicks. The change is in how you write them. Answer the common question briefly and accurately at the top. Then go where summaries cannot: your data, your cases, your judgement, your tools. A post on "how to choose accounting software" that ends with a scored comparison of eight products, tested on the same sample company, earns citations and clicks. A post that lists "consider your budget, check integrations, read reviews" earns neither.

How do you measure whether a topic still earns its place?

• Clicks from search, by post type, over time.
• Citations in AI answers for each post's main question, checked quarterly.
• Conversions or sign-ups attributed to each post.
• Links and mentions the post attracts.

A post with no clicks can still be worth keeping if it is regularly cited, especially on topics close to buying decisions. A post with none of these is a candidate to upgrade or retire.

What might an illustrative case look like?

Consider an illustrative accounting software company with 400 blog posts, most of them definitions and tip lists written for keywords. Search clicks to those posts had halved in a year, and AI Overviews answered most of their questions in full. The team merged 120 thin definition posts into a glossary and retired 80 tip lists with no links or traffic. It upgraded 40 how-to posts with screenshots, sample files and troubleshooting sections. It started a quarterly survey of small business bookkeeping habits. Six months later, total blog traffic was lower, but sign-ups from the blog had risen. AI answers were citing the survey and several upgraded guides.

What do people ask about blog topics and AI summaries?

Is SEO blogging dead?

No, but blogging for keywords alone is fading. Posts that add original information still earn citations and clicks.

Should we delete old posts that lost traffic?

Only if they have no links, no citations and no conversions. Otherwise upgrade or merge them.

Do AI summaries use our posts even when people do not click?

Often, yes. That builds recognition, which matters for later recommendation questions.

Which new topics should we prioritise?

Those where you have data, experience or tools others lack, and those closest to your buyers' decisions.