m
Manifest Start free trial
Pillar guide · April 2026

AI tools for newsrooms: the working editor's playbook for 2026

Half the AI-tools-for-newsrooms posts on the open web are written by people who have never sat in an editorial meeting. This one is written by a team that spends every working day inside publisher archives. Below: the six free tools we use ourselves, the order to run them in, and the cheapest changes that move the needle on AI-search visibility this quarter.

The shape of the problem

ChatGPT, Perplexity, Claude, and Google's AI Overviews are quietly becoming the front page of the web for an enormous and growing share of news queries. For most newsrooms the per-engine pattern looks the same:

  • The AI engines do reach the homepage. They struggle on the article pages where the actual reporting lives.
  • The article pages are missing the structured signals (NewsArticle JSON-LD with speakable, dateline, byline, and isAccessibleForFree) those engines use to decide what to quote.
  • Reporters don't have author bio pages with Person schema, so AI engines can't ground a citation in a verifiable identity.
  • There's no llms.txt, so the engines have to guess the shape of the archive.
  • The headlines are written for the homepage, not for an answer engine.
  • And whole beats with high search demand — school board meetings, zoning, restaurant inspection scores — aren't covered, so the engines invent answers.

Each of those is fixable in less than a day of editorial-engineering time. Here are the six free tools we built to do it.

1. Run an AI-visibility audit

Start with the AI visibility checker. Paste your domain. You get per-engine scores for ChatGPT, Perplexity, Claude, and Google AI Overviews, plus a prioritized fix list. We've also pre-run the audit for ~120 well-known US publications — e.g. The New York Times, The Texas Tribune, Block Club Chicago — so you can compare against your peers without a signup.

2. Publish a publisher-shaped llms.txt

Use the llms.txt generator. Feed it your domain, your beats, and your roster. It produces a publisher-shaped llms.txt that lists your beats, your reporters, your archive sections, and your citation rules. Drop it at the root of your site. Re-generate any time the roster changes.

3. Add NewsArticle JSON-LD to every story

The single highest-leverage change you can make this week is to ship valid NewsArticle structured data on every article page, with speakable, dateline, isAccessibleForFree, and a real mainEntityOfPage ID. Use the NewsArticle schema generator to produce a copy-paste snippet your CMS can template.

4. Rewrite headlines for SEO + GEO

AI engines reward declarative, fact-forward headlines that lead with the news. The headline rewriter takes any headline and returns five rewrites optimized for both classic Google SEO and generative engine citation, with reasoning per variant. We use it on every wire story before it gets pinned.

5. Ship a real reporter bio page for every byline

Author pages are the most-overlooked SEO and AI-visibility lift in the entire newsroom stack. The reporter bio generator writes an SEO-clean bio, a title tag, a meta description, and a Person JSON-LD block in one click. Ship one for every reporter on staff this quarter.

6. Find the beats you're not covering yet

Use the local-news topic gap finder to surface the beats with high search demand and low publisher coverage in your city. These are the gaps the AI engines are filling with bad answers right now — school board, zoning, court arraignments, inspection scores, hyperlocal real-estate briefs. Each one publishes evergreen long-tail traffic.

What to do this week

  1. Run the AI-visibility checker. Save the report.
  2. Ship NewsArticle JSON-LD across the article template. (Highest leverage.)
  3. Publish your llms.txt.
  4. Generate one reporter bio page; ship the rest by Friday.
  5. Re-run the audit Monday. The score will move.

Where Manifest fits in

These free tools are samples of what Manifest does live, on every visit, on autopilot. Mani is an on-site AI assistant trained on your archive that answers reader questions in your voice, captures the lead, and turns the readers AI engines refer to you into a relationship you own — not a one-time citation in someone else's product.

Start a 7-day free trial →  $20/mo after trial. No card to start.