Publishing many pages with AI, without a spam problem

AI makes it possible to publish a hundred pages in a day. This chapter explains where Google draws the line, how location and service pages cross it, and how to plan a steady pace with a reason for every page.

Google does not set a page limit. Its scaled content abuse policy targets many pages made "for the primary purpose of manipulating search rankings and not helping users", however they are produced. Publish a page only when you can name the customer question it answers and the facts it adds that other pages lack. Pages that differ only by a town name, a keyword or a reworded source are the pattern the policy describes.

Volume itself is allowed

Google's policies set no limit on the size of a site, and its 2023 guidance points out that automation has long produced useful content "such as sports scores, weather forecasts, and transcripts." A weather page for every town is useful because each one carries different data that people need.

What Google's spam policy describes is something else: "when many pages are generated for the primary purpose of manipulating search rankings and not helping users." The words that matter are "primary purpose" and "not helping users". In March 2024 Google said the policy now applies "whether automation, humans or a combination are involved", so the question is the same for a team of writers and for a chatbot: does each page help someone?

The five patterns Google names

Google's spam policy lists five examples of scaled content abuse. Each one is a shortcut that AI makes faster, which is why they are worth knowing before a business plans a large publishing project.

Google's rater guidelines add one sentence that summarises the problem: "Creating an abundance of content with little effort or originality with no editing or manual curation is often the defining attribute of spammy websites."

• Using generative AI tools "to generate many pages without adding value for users". The test is the added value, page by page.
• Scraping feeds, search results or other content to generate many pages, including through "synonymizing, translating, or other obfuscation techniques". Rewording another site's page with AI is one of these techniques.
• "Stitching or combining content from different web pages without adding value". A page that merges the top five search results into one is the typical case.
• "Creating multiple sites with the intent of hiding the scaled nature of the content."
• "Creating many pages where the content makes little or no sense to a reader but contains search keywords."

Location pages and service pages

The most common place where a small business meets this problem is the location page. A plumber serving twelve towns wants to appear when someone searches "plumber" with each town's name. The quick solution is twelve copies of the same page with the town name changed, now easy to produce with AI in minutes. This is the pattern the policy describes as many pages without added value.

A location page earns its place when it carries facts specific to that place. For a trades business, that can mean the jobs completed there, with photos, the local permit office and its rules, the type of housing common in the area and the problems it causes, the travel time and any extra charge, and reviews from customers who live there. If the business cannot find three or four such facts for a town, a single page listing all the areas it serves is the honest choice.

Service pages follow the same logic. One page for each service the business offers is useful. Twenty pages for small variations of the same service, each with similar text, are the opposite. Each service page should describe what the work involves, what it costs or how it is priced, how long it takes and what the business has done of this kind.

A location page, built the right way

Here is an outline written for this guide, for an imaginary electrician that serves several towns. It shows what a location page can hold when the business collects real facts for it first.

Every item on that list comes from the business's own work. No chatbot can produce them, and no competitor can copy them without the copy being obvious. A town where the business has no jobs, photos or reviews yet belongs on the general list of areas served until the business has worked there.

• A first paragraph that answers the search directly: the business works in this town, how soon it can usually arrive, and how to book.
• Three recent jobs in the town, each with a photo, the type of house, the problem found and how it was solved.
• The local rules the electrician meets: which office issues electrical permits for that town, and what inspection it requires.
• The common problems in the local housing, taken from the team's own records, for example older wiring in houses from a given decade.
• Reviews from customers who live in the town, used with their permission.
• A short list of questions that customers in the town have asked, with the answers.

How many pages a week is safe?

Google gives no number, and anyone who quotes one is guessing. Its help page on people-first content warns against "producing lots of content on many different topics in hopes that some of it might perform well in search results", and against "using extensive automation to produce content on many topics." It also asks whether a site has "a primary purpose or focus".

For a small business the right pace is the pace at which every page can pass the editing process in chapter 4: a brief, a fact list, first-hand details, a full edit, a read-aloud and a named person's approval. For some businesses that is one page a week. For a business with a dedicated writer and an owner who answers questions quickly, it can be more. The constraint is the checking, and the pace follows from it.

It also helps to stay on topic. A roofing company publishing articles about roofs, gutters and insulation builds a site with a clear focus. The same company publishing articles about mortgages, interior design and travel, because a tool suggested those keywords, drifts away from the "primary purpose or focus" that Google's question describes, and away from the expertise readers expect from a roofer.

A plan for publishing with AI at a steady pace

The plan below suits a small business that wants to publish regularly with AI help, while keeping every page worth reading.

• Start from questions. Collect the questions customers ask, from the phone, from emails, from the search terms in Google Search Console. Each question that the site does not answer yet is a candidate page.
• Group the questions. Questions that a single page can answer together become one page. This prevents the near-duplicate pages that the policy targets.
• Order by value. Put first the questions that come before a purchase: cost, timing, choosing a provider, local rules.
• Gather the facts for each page before writing, as chapter 6 describes.
• Write and edit each page through the eight steps in chapter 4. A page that cannot get first-hand details waits until it can.
• Link the pages together: each article to the service it relates to, each service to its articles. A reader who arrives on one page should find the next one easily.
• Review every month. Look at which pages bring visits and enquiries in Google Search Console, update the ones whose facts changed, and merge or remove pages that answer nobody's question.

Cleaning up after a period of mass publishing

Some businesses arrive at this guide after publishing many AI pages already. Google's spam policy gives the direction: if you host content made at scale without value, "exclude it from Search". The work has three parts.

First, list every page and sort them by visits from Google over the last twelve months, using Search Console. Second, read the pages with no visits and decide, page by page, whether each one answers a real question. Third, act: improve the pages that answer a real question with the process in chapter 4, merge pages that cover the same question into the strongest one and send visitors from the old addresses to it, and remove the rest.

Removing pages can feel like losing work. Pages that bring no visits and help no reader give the business nothing, and the pages that remain are the ones a customer will judge it by.

What scale looks like when it works

Part of a larger guide

This page is one part of Avoiding AI slop. The other parts:

• What AI slop is
• What Google says
• Signs of AI writing
• The editing process
• First-hand experience
• Checking facts
• Bylines and disclosure

Questions people ask

Is there a limit to how many AI pages I can publish?

Google sets no number. Its scaled content abuse policy targets many pages generated "for the primary purpose of manipulating search rankings and not helping users", however they are made. A site can hold thousands of useful pages. The question for each page is whether it helps a reader with facts or value that other pages do not already provide.

Are location pages considered spam by Google?

Location pages are fine when each one carries facts specific to the place: jobs completed there, local rules, local photos, reviews from local customers. Copies of the same page with only the town name changed fit the pattern Google's policy describes as many pages without added value. If you cannot find several real local facts for a town, list it on one areas-served page instead.

How many blog posts should a small business publish per week?

As many as it can check properly, and Google gives no target. Its help page lists "producing lots of content on many different topics in hopes that some of it might perform well" as a warning sign. If each post can go through a brief, a fact check, first-hand details and a named approval, the pace is right. If not, publish less.

Is rewording other websites' content with AI allowed?

Rewording to create many pages is one of the examples in Google's scaled content abuse policy, which names "synonymizing, translating, or other obfuscation techniques" used on scraped content where little value is added. Google's rater guidelines also tell raters to give the lowest rating to pages that are mostly paraphrased with little effort or originality. Add your own facts and analysis instead.

What should I do with old AI pages that get no visits?

Sort all pages by visits from Google over the last twelve months in Search Console. Read the ones with none and decide page by page: improve those that answer a real customer question, merge near-duplicates into the strongest page and redirect the old addresses, and remove the rest. Google's spam policy says content made at scale without value should be excluded from Search.

Does publishing on many unrelated topics hurt a site?

Google's help page asks whether a site has "a primary purpose or focus", and lists writing about topics "simply because they seem trending" or entering a niche "without any real expertise" as warning signs. A business site is strongest on the subjects its team knows from its own work. Articles far from that work also attract readers who are unlikely to become customers.

Can a team of human writers also break the scaled content policy?

Yes. When Google widened the policy in March 2024, it said it would focus on "producing content at scale to boost search ranking", "whether automation, humans or a combination are involved." Hiring writers to produce hundreds of thin pages is treated the same as generating them with AI. What the policy weighs is the value of each page.

How do I decide which page to write next?

Start from real questions customers ask, collected from calls, emails and the search terms in Google Search Console. Group questions one page can answer together, then write first the ones asked before a purchase: cost, timing, choosing a provider and local rules. Each new page should answer a question the site does not answer yet.

Is programmatic content from my own data acceptable?

It can be, when each page carries useful data people need. Google's 2023 guidance notes that automation has long produced helpful content "such as sports scores, weather forecasts, and transcripts." A page per product built from your real specifications, or a page per area built from real local records, helps readers. Pages built from a template with no distinct data do not.

Should I keep publishing while I clean up old content?

Clean up first, or at least at the same time. Pages already published are the ones visitors meet, and a site with many weak pages and a few new good ones still shows readers mostly weak pages. Fix or remove the pages with no value, then return to a steady pace where every new page passes the full editing process.

Sources

1. Spam policies for Google web search: scaled content abuse, Google Search Central. Last updated 2026-08-28, accessed 4 November 2026.
2. New ways we're tackling spammy, low-quality content on Search, Google, The Keyword. Published 5 March 2024, updated 26 April 2024, accessed 4 November 2026.
3. Google Search's guidance about AI-generated content, Google Search Central Blog. Published 8 February 2023, accessed 4 November 2026.
4. Creating helpful, reliable, people-first content, Google Search Central. Last updated 2026-10-05, accessed 4 November 2026.
5. Search Quality Evaluator Guidelines, Google. Version of 11 September 2025, accessed 4 November 2026.

Keep reading

• How to avoid AI slop: the full guide
• What Google says about content written with AI
• What AI slop is, and how to spot it on your site
• The editing process from AI draft to published page
• Small business SEO, explained from the first step
• Run the free Get Found Check

Want pages that sound like your business and bring enquiries?

Licheo's agents write and update your pages from your own facts, and a person checks every page before it goes live. Start with the free Get Found Check to see how customers find you today, or contact us for pricing.