We publish 20 blog posts a month and none get cited by ChatGPT or Perplexity: what are we doing wrong?

Volume is rarely the problem

Publishing 20 posts a month and getting no citations from ChatGPT or Perplexity is a common situation, and the cause is almost never that you need to publish more. Assistants do not reward cadence. They retrieve pages that match a specific question, can be fetched cleanly, and offer something worth quoting. If none of your posts are being picked, one of those three conditions is probably failing, and you can find out which with data instead of guessing.

We are not going to cite a study to tell you what "AI search favors". We have no source we can check, and a confident number would not help you anyway. Here is a diagnosis order that works on facts you can pull from your own site.

Step 1: Check whether the bots can read your posts

This is the first check because it is binary. If the AI crawlers cannot fetch a page, no amount of editing matters.

Is robots.txt blocking bots you want, such as OAI-SearchBot, ChatGPT-User, PerplexityBot or Perplexity-User?
Is a CDN, WAF or bot-protection rule refusing them even though robots.txt allows them?
Do pages load without JavaScript, or do bots receive an empty shell?
Are the bots getting errors, redirects or timeouts on your blog URLs?

Most teams find out only when someone looks at server or CDN logs. Promptwatch publishes a customer story that fits: TreeHouse, which runs Pararius and Huurwoningen.nl, reports that AI search became about 25% of its organic conversions, 99% of it from ChatGPT, after fixing crawler access and automating weekly reporting with the Promptwatch MCP. That is TreeHouse's own result, not something to expect by default, but the sequence (access first, then measurement) is sound.

Step 2: Check whether your posts match real prompts

A blog calendar built from keyword tools can produce posts nobody asks an assistant about. Assistants break a question into narrower searches (query fan-outs), then retrieve pages matching those. A post titled around an internal topic may not match any of them.

Pull the prompts your buyers ask, look at the fan-outs the models generate, and compare them with your post titles. The mismatch is often obvious in an hour.

Step 3: Check whether the posts say anything new

Twenty posts a month at a steady cadence often means twenty posts that restate what the top ten results already say. An assistant choosing sources has no reason to quote a page that adds nothing. Things that give a page a reason to be cited:

Original data, even small: your own customer questions, internal benchmarks you can stand behind, results you can name.
A plain, direct answer in the first paragraph.
Named sources and dated facts.
A clear point of view on a comparison, with criteria.
Specifics a competitor cannot copy: your product's real behavior, your process, your pricing logic.

Step 4: Check whether anyone else mentions you

Assistants lean on third-party pages. If your brand appears on no review sites, forums or comparison articles, your blog is speaking alone. Look at where competitors are named and you are not: Reddit threads, YouTube videos, roundups.

Step 5: Check that you are measuring the right thing

You may be getting cited and not seeing it, because you are not tracking the relevant prompts. Or the posts may be fetched and read but never cited. Those are different problems with different fixes.

Tools that find the cause

Promptwatch is the platform we would start with because it connects each of those checks. Agent Analytics provides real-time, IP-verified AI crawler logs for bots including GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot and PerplexityBot, with a crawl-to-citation path, a citation rate per page and error tracking with failure explanations. Query fan-outs per model show what assistants actually search. Content gap analysis scores your coverage against the prompts you track, so you see which topics lack a good page. Citation analytics and offsite mentions show who gets cited instead of you and where your brand is named without a link. Page Tracker follows specific URLs, including a competitor's post that keeps winning. The Action Board then produces a prioritized list of evidence-backed actions.

On the creation side, the Content Agent (Business and Enterprise) plans, writes and publishes against visibility gaps on a schedule, with a review inbox or a hands-off mode with a quality check. Crisp, a Promptwatch customer, reports that AI-referred traffic converted at twice the rate of traditional channels and that it scaled to 5 to 10 articles a day. That is Crisp's published case, not a general expectation. Crawler logs begin on Professional ($245 a month), and the Content Agent on Business ($579 a month).

AthenaHQ is a sound choice if you want an action queue and a content optimization agent. Starter is $295 a month, credit-based. Its catalog entry lists no crawler-log product, so Step 1 stays manual.

Scrunch AI includes AI agent and bot traffic reporting, with Core at $250 a month, and its Agent Experience Platform serves AI-optimized pages to crawlers at the Enterprise tier.

Gauge includes a content engine that writes publish-ready articles: Growth is $599 a month for 18 articles, 600 daily prompts and three platforms.

Profound sells marketing agents for content and PR, with Enterprise contracts starting around $40,000 a year per our existing entry.

A 30-day plan

Week 1. Audit access. Fix robots.txt, CDN or rendering problems. Install visit tracking.
Week 2. Build a prompt set of 30 to 50 buyer questions. Check which posts, if any, match them.
Week 3. Pick the five gaps with the most buyer relevance. Replace volume with depth: five strong, original pages beat twenty generic ones.
Week 4. Re-measure. Look at crawler fetches, citations and AI-referred visits for those pages.

Cut the calendar if you have to. Fewer, better posts are cheaper and more likely to be cited.

What nobody can promise

No tool or tactic guarantees a citation. Engines change, answers vary between askings, and competitors publish too. What the diagnosis gives you is a reason, and a reason is what you were missing.

Who should pick which

You need the diagnosis end to end (access, prompts, gaps, citations, traffic) plus content: Promptwatch.
You want an action queue and content optimization, and can check crawling separately: AthenaHQ.
You want bot reporting and crawler-side serving: Scrunch AI.
You want a fixed monthly article quota alongside tracking: Gauge.