Integrality — NotesField notes on data pipelines, AI agents in production, and software for regulated environments.https://integrality.io/AI travel agents have a data problem before they have an AI problemhttps://integrality.io/blog/ai-travel-agents-data-layer/https://integrality.io/blog/ai-travel-agents-data-layer/Better models do not solve stale prices, missing availability or fragmented travel catalogs. AI travel agents need a fresh, structured data layer underneath.Fri, 14 Aug 2026 00:00:00 GMTModern aviation data pipelines: APIs, NDC and the gaps between themhttps://integrality.io/blog/modern-aviation-data-pipelines/https://integrality.io/blog/modern-aviation-data-pipelines/Modern travel products combine schedules, status, airline retail content and external sources. Here is where APIs, NDC and targeted collection each fit.Fri, 14 Aug 2026 00:00:00 GMTWhy travel activities data is an entity-resolution problem, not a scraping problemhttps://integrality.io/blog/travel-activities-data-normalization/https://integrality.io/blog/travel-activities-data-normalization/Tours and activities are fragmented across operators, marketplaces and reservation systems. The hard problem is mapping them into one reliable, fresh catalog.Fri, 14 Aug 2026 00:00:00 GMTTravel data collection at scale: the scraper is the easy parthttps://integrality.io/blog/travel-data-collection-at-scale/https://integrality.io/blog/travel-data-collection-at-scale/Building a travel scraper is easy. Keeping airline, hotel and activity data complete, fresh and reliable as sources change is the real engineering problem.Fri, 14 Aug 2026 00:00:00 GMTHTTP 200 is not uptime: the silent failures that break travel data pipelineshttps://integrality.io/blog/travel-data-pipeline-monitoring-http-200/https://integrality.io/blog/travel-data-pipeline-monitoring-http-200/A travel data pipeline can be technically online while silently losing coverage, freshness or critical fields. Here is how to monitor the data, not only the server.Fri, 14 Aug 2026 00:00:00 GMTWhy scrapers die in month threehttps://integrality.io/blog/why-scrapers-die-in-month-three/https://integrality.io/blog/why-scrapers-die-in-month-three/Most scraping projects work perfectly for about eight weeks. What happens next is predictable, and it has almost nothing to do with how the scraper was written.Tue, 04 Aug 2026 00:00:00 GMTYour AI pilot didn't fail on the modelhttps://integrality.io/blog/your-ai-pilot-didnt-fail-on-the-model/https://integrality.io/blog/your-ai-pilot-didnt-fail-on-the-model/AI pilots rarely fail because of model choice. The real problem is usually stale, fragmented or poorly governed data underneath the production system.Tue, 21 Jul 2026 00:00:00 GMTNobody asks about the audit log during a demohttps://integrality.io/blog/nobody-asks-about-the-audit-log/https://integrality.io/blog/nobody-asks-about-the-audit-log/Access control and audit trails are the least demonstrable parts of regulated software, and the hardest to add later. A note on building them first.Tue, 30 Jun 2026 00:00:00 GMT