NewsBreak blocks AI crawlers everywhere except its local business pages, and llms.txt tells assistants to use the local-business sitemap. On October 21, 2026, 59 of the 200 most recently updated entries in that sitemap returned "Page not found". The listings that do load share one generic news description, carry no LocalBusiness markup, and follow a /local/spot/<id> pattern that llms.txt does not describe. The new local ads product manager is hired to grow discovery through business profiles. Right now the one surface NewsBreak opens to AI search is sending assistants to dead pages.
First action
Export today's sitemap, request every URL, and list the ones that fail.
How to measure it
Live sitemap entries
- Now
- –
- Target
- 99.5%+
- By
- 30 days
Steps
Measure
Status of all 1,801 entries, grouped by last-modified date.
Web engineering 1 day
Find the cause
Why recently updated listings disappear: merges, removals or a publishing bug.
Web engineering 2 days
Filter the sitemap
Only listings that return 200, all on https.
Web engineering 2 days
Redirect removed listings
To the business's new page or the city and category page.
Web engineering 2 days
Unique descriptions
Business, category and city in every title and description.
Senior Product Manager, Local Ads and Monetization 1 week
Structured data
LocalBusiness markup with address, phone, hours and rating.
Web engineering 1 week
Fix llms.txt
Describe the URL patterns actually in use.
Senior Product Manager, Local Ads and Monetization 1 hour
Monitor
Nightly check with an alert above 1% failures.
Web engineering nightly
Why it works
Generate the sitemap only from listings that return 200, and drop or redirect removed ones within a day. Give every listing its own title and meta description (business, category, city), add LocalBusiness structured data (name, address, phone, hours, rating), and update llms.txt so the URL patterns it describes match the real ones. Serve every sitemap entry on https. Then add a nightly check that alerts when more than 1% of entries fail.
Search engines and AI assistants stop trusting a sitemap that sends them to dead pages, and unique, marked-up listings are what they quote when someone asks "who is a good plumber in Missoula". Every listing ends with "Claim this listing", so more pages found means more owners claiming them and more owners in the funnel to buy local ads.
Ready to use
ChecklistLocal listing page checklist
[ ] Returns 200 over https, or is out of the sitemap [ ] Title: business, category, city [ ] Meta description written for this business [ ] LocalBusiness JSON-LD (name, address, phone, hours, rating) [ ] Canonical matches the sitemap URL [ ] URL pattern matches llms.txt [ ] "Claim this listing" call to action [ ] Home-services listings link to Nearby AI
Risks and how to handle them
Listings removed for good reasons (closed businesses)
Return 410 for closed businesses and keep them out of the sitemap.
Thin listings indexed
Keep listings with no hours or reviews out of the sitemap until they are complete.
Evidence
- Your websiteThe local-business sitemap that llms.txt points AI systems to lists 1,801 pages, all under /local/spot/<id> rather than the /local/{city-state}/ pattern llms.txt describes. Checked October 21, 2026: 59 of the 200 most recently updated entries returned "Page not found" (HTTP 404); samples from the middle and the end of the list loaded. The second local sitemap (local-business.xml) lists three URLs, one of them over http rather than https.Source
- Your websiterobots.txt blocks about 35 AI crawlers (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended and others) from everything except /local and the local-business sitemaps. llms.txt calls local business pages "the only surface NewsBreak makes available to AI systems", and says a business's canonical page is "localbusiness.newsbreak.com/{city-state}/{business-slug}", falling back to "www.newsbreak.com/local/{city-state}/{business-slug}".Source
- Your websiteA sampled listing (a plumber in Missoula, MT) shows rating, review count, phone, address and hours, then "This intro was compiled from public data ... Are you the owner? Claim this listing to add hours, photos, and offers." Its meta description is the site-wide "The best local & breaking news source in the US..." text, it carries no LocalBusiness structured data, and Nearby AI appears only as a footer link.Source
- HiringSenior Product Manager, Local Ads and Monetization (posted October 14, 2026): owns "local ads and local services", builds "experiences that help users discover relevant local businesses ... through sponsored content, business profiles, recommendations", must "simplify business onboarding, campaign creation, geographic targeting" and improve "activation, repeat spending, and retention across self-service and managed accounts", and launch "AI capabilities that improve local discovery".Source
- HiringHead of Growth & Go-to-Market, Nearby AI (posted September 11, 2026): "A new consumer marketplace has to be found before it can be trusted." "Our most defensible growth asset is proprietary, structured local data that no incumbent can publish. Turning that data into content that search engines reward and AI assistants cite is the center of this role." Owns a "programmatic SEO program on proprietary structured data, with editorial standards", "AI-search visibility", lifecycle marketing "within the company's consent principles" and the public brand launch. Principles: "no pay-for-rank", no sharing contact details beyond what the customer approved, "no quality claims we cannot substantiate".Source