{"slug": "improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage", "title": "Improve AI Output with a News Feed: Ground LLM Answers in Live Coverage", "summary": "Perigon has introduced a news feed API designed to ground large language model outputs in live, structured coverage, addressing issues like stale knowledge and hallucinated citations. The service offers entity-pinned articles and stories that developers can retrieve and pack into prompts as evidence, with filters for companies, people, and domains to keep context clean. A worked example shows how to use companyId and personWikidataId to track Stripe and its CEO Patrick Collison.", "body_md": "Teams ship chatbots, research agents, and support copilots that sound confident on last week’s events - then invent partnerships, mix up CEOs, or cite outlets that never ran the story. Model weights are stale. A generic web scrape is noisy. What fixes **AI output** for current events is a **news feed for AI**: structured, filterable coverage you retrieve first, then pass into the prompt as evidence.\n\nFintech and payments desks see this clearly. A model asked “what is Stripe doing in sponsorships?” without a feed will improvise. The same question with a **grounded AI** context pack - recent articles and story clusters pinned to the company - stays on published facts.\n\nDevelopers usually ask:\n\nWhy does raw LLM output fail on live news without a feed?\n\nHow do you build a filtered news feed as LLM context?\n\nHow do you pack article and story fields into prompts without drowning the model?\n\nWhen should you use Articles, Stories, Search Summarizer, or MCP?\n\nWhich filters keep a company or executive context pack clean?\n\nThe sections below answer each in order.\n\nTry it live:[Articles sandbox]·Read the docs:[Entities]·Compare plans:[API pricing]\n\nAn LLM without retrieval answers from training cutoffs and pattern completion. That fails when the user needs **live news for AI**:\n\n**Stale knowledge.** Product launches, partnerships, and regulatory moves after the cutoff never appear - or appear as confident fiction.\n\n**Homonym collisions.** “Stripe” and “Collison” hit sports, entertainment, and unrelated brands unless the feed uses entity IDs.\n\n**No citations.** Fluent prose without publisher, date, and title makes compliance and desk review impossible.\n\n**Unscoped volume.** Dumping the open web into the context window adds forums, reprints, and off-topic inflation pieces that dilute the answer.\n\nA **RAG news feed** (retrieve → pack → generate) keeps the model honest: the corpus is your filtered **news data for AI**, not the entire internet.\n\nTreat Perigon Articles (/v1/all) as the retrieval layer for **LLM context news**. Resolve the company once, then poll a tight window.\n\nWorked example - payments company watch:\n\nRole |\nPrefer |\nValue |\n|---|---|---|\n| Stripe (payments) | companyId |\n9c0e5ff08b87453b8f6fa108a73b57b5 |\n| Stripe (domain pin) | companyDomain |\nstripe.com when you want domain-linked hits |\n| Patrick Collison (CEO) | personWikidataId |\nQ7146257 |\n\nOn Articles, pass **companyId** set to 9c0e5ff08b87453b8f6fa108a73b57b5, set **sortBy** to date, and bound **from** / **to** (or **addDateFrom** for incremental polls). Keep **size** small for a context pack - often 5 to 15 articles - so the model can read titles and summaries without truncating mid-sentence.\n\nPeople search for “Patrick Collison” can return basketball and entertainment Collisons. Keep **personWikidataId** Q7146257 when the lane is leadership commentary, not the company-wide feed. Deep entity param theory lives in the Company News API guide; this page owns the **improve AI output** packing pattern.\n\nFor narrative heat without reprint spam, run the same **companyId** on Stories (/v1/stories/all). Coverage often looks like a sponsorship or partnership cluster - for example a Stripe worldwide partner announcement spanning multiple outlets under one story name - which compresses better into a prompt than eleven near-duplicate headlines.\n\nRetrieval is not enough. **Ground LLM** quality depends on what you paste and what you instruct.\n\n**Context pack recipe (per hit):**\n\nTitle\n\nPublisher domain\n\nPublication date\n\nShort summary or key points (not full body by default)\n\nOptional company / person names already on the record\n\n**Prompt scaffolding:**\n\nInstruct the model to answer only from the packed items.\n\nRequire citations as title + domain + date.\n\nCap length (bullets or a short brief).\n\nSay what to do when the pack is empty: admit gaps, do not invent.\n\n**Size discipline:** Prefer recent, entity-pinned hits over a large unsorted dump. Default article order without **sortBy** date can surface old or off-beat pages first - useless for **AI with live news**.\n\nRaise quality with **sourceGroup** (for example top100) and **excludeLabel** Opinion, Paid News, or Press Release when you want third-party reporting only.\n\nSearch Summarizer can turn the same filters into one model-written briefing when the deliverable is a digest, not a custom agent reply - see the [AI News Summaries guide](https://perigon.io/blog/ai-news-summaries-api-turn-news-search-results-into-executive-briefings). Keep this article on feeding **your** LLM.\n\nJob |\nPrefer |\nWhy |\n|---|---|---|\n| Custom chatbot / agent answers with citations |\nArticles → your LLM |\nYou own the prompt, tools, and UX |\n| Fewer reprints in the context window |\nStories first, then sample articles |\nOne narrative per cluster |\n| One leadership-ready digest, Perigon-written | Search Summarizer |\nPrompted briefing of the matched set |\n| IDE / assistant tool access without custom retrieval code | MCP |\nAgent tools over the same news graph - see\n|\n\n**News feed grounding** and Summarizer both start from filtered search. Grounding keeps generation in your stack; Summarizer returns the briefing as the API response.\n\nGroup parameters by the grounding job:\n\nUse case |\nParams to teach first |\n|---|---|\n| Pin the company |\ncompanyId, optional companyDomain\n|\n| Pin an executive | personWikidataId |\n| Freshness for RAG |\nfrom / to, addDateFrom, sortBy date |\n| Context window budget |\nsize, page\n|\n| Publisher quality |\nsource, sourceGroup\n|\n| Cut non-reporting noise | excludeLabel |\n| Optional topic fence |\ncategory, topic, short q\n|\n\nStart with **companyId** for Stripe-scale watches. Add **personWikidataId** only for CEO interview or leadership lanes - company volume and person volume differ sharply (see facts below). Stack **excludeLabel** and a trusted **sourceGroup** before you grow **size**.\n\nEach Articles hit gives the fields most agents need: title, url (when your product deep-links), pubDate, publisher domain, summary, and linked companies or people. You map those into the context block; you do not need full HTML.\n\nStories return a cluster name, summary, article count, and update time - enough to tell the model “this is one partnership narrative with N outlets,” then attach one or two representative article rows if citations matter.\n\nPattern check (evergreen shape, not a permanent newsjack): a recent Stripe-linked story can look like a multi-outlet sports-partnership announcement (Ryder Cup worldwide partner framing) with on the order of ten related articles under one cluster - ideal Stories-first packing when reprints would otherwise fill the window.\n\nRetrieve the last week of **companyId** coverage before the model answers “what changed for this vendor?” Pack titles and domains so agents cite press, not rumor.\n\nRun parallel packs: company feed vs **personWikidataId** leadership feed. Compare volumes, then ask the model for a short memo with separate citation lists.\n\nShip a “sources used” panel from the same pack you sent to the model - title, domain, date - so users can audit **grounded AI answers**.\n\nGoal |\nPath |\n|---|---|\n| Improve answers inside your LLM app | Articles/Stories → context pack → your model |\n| One digest without building a packer | Search Summarizer |\n| Always-on desk briefing product | Signals on perigon.io/home (product path; not this API recipe) |\n| Entity filter depth | Company News API guide |\n\nEntity pins set who. Dates and **sortBy** set when. **sourceGroup** and **excludeLabel** set quality. Optional **q** (partnership, pricing, regulation) narrows event language without replacing the allowlist. Poll with **addDateFrom** so each agent turn only spends tokens on new hits.\n\nPayments coverage is continuous: product, partnerships, regulation, and executive commentary move on different cadences. Entity pins keep a **news feed for AI** on the company instead of every “stripe” metaphor in sports or design.\n\n**What that looks like in the index (30 days, pubDate 2026-06-23 to 2026-07-23):** about **11,000** articles linked to Stripe (**companyId** 9c0e5ff08b87453b8f6fa108a73b57b5).\n\nLeadership is a thinner lane: about **150** articles for Patrick Collison (**personWikidataId** Q7146257) in the same window. Pack company and CEO feeds separately so a support copilot does not drown in executive profile pieces - or miss a partnership cluster when it only watched the person ID.\n\n**Improve AI output** on current events by retrieving a filtered news feed before generation - not by hoping the model remembers.\n\nPin entities (**companyId**, **personWikidataId**), sort by date, and keep **size** small enough for reliable citations.\n\nPack title, domain, date, and short summary; instruct the model to stay inside the pack.\n\nUse Stories to compress reprints; use Search Summarizer or MCP when you want a Perigon briefing or agent tools instead of a custom packer.\n\nSeparate company-scale and executive-scale watches - volumes differ, and mixed packs confuse the model.\n\nGet started:[Entities]·[API reference]·[Articles sandbox]·[API pricing]", "url": "https://wpnews.pro/news/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage", "canonical_source": "https://dev.to/alexperigon/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage-81i", "published_at": "2026-08-21 10:25:48+00:00", "updated_at": "2026-08-21 10:44:51.910162+00:00", "lang": "en", "topics": ["large-language-models", "ai-products", "developer-tools", "ai-infrastructure"], "entities": ["Perigon", "Stripe", "Patrick Collison"], "alternates": {"html": "https://wpnews.pro/news/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage", "markdown": "https://wpnews.pro/news/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage.md", "text": "https://wpnews.pro/news/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage.txt", "jsonld": "https://wpnews.pro/news/improve-ai-output-with-a-news-feed-ground-llm-answers-in-live-coverage.jsonld"}}