News Research

Description

Research any topic across recent news sources — with sentiment analysis, source diversity, and professional reporting. Supports single topics, comparative analysis (e.g., "Tesla vs BYD"), and multi-entity research. Reports include per-article sentiment classification, executive summaries, and full source citations.

Core Capabilities

• Single topic, comparative ("X vs Y"), and multi-entity research in one agent
• Per-article sentiment classification (positive, negative, neutral, mixed) with overall sentiment label
• Quality controls: minimum 8 articles, 6+ unique sources, relevance scoring
• Multi-language support
• Configurable date range and focus area instructions

Good Query Examples

• "Latest news about Apple's AI strategy from the past 7 days" (cutoff_date: today)
• "Tesla vs BYD in electric vehicles — coverage over the last 30 days" (cutoff_date: today)
• "How did the US tariff situation develop month over month during 2025?" (cutoff_date: 2025-12-31)
• "French media sentiment toward US tech companies throughout 2024" (cutoff_date: 2024-12-31)

Bad Query Examples

• "What do consumers think about electric vehicles?" → This asks for consumer attitudes, not news coverage. Fix: Use GWI Spark for consumer behavior data.
• "Get me the full Financial Times analysis on the EU AI Act" → This agent cannot access paywalled content. Fix: Broaden the topic — "EU AI Act coverage over the past 30 days" will pull from publicly available sources.
• "What will happen to AI stocks next quarter?" → This agent analyzes existing news coverage, it cannot predict or forecast. Fix: Frame as "Recent news coverage about AI stock performance from the past 30 days."
• "News about AI from the past 7 days" (cutoff_date: 2026-01-01) → The cutoff date contradicts the query — you're asking for the last 7 days but set the cutoff to over a month ago. Fix: Set cutoff_date to today if you want recent news, or adjust your query to match the date range.
• "News about electric cars" → Too vague, no time period specified. Fix: Be specific about the topic and timeframe: "News about EV battery technology breakthroughs over the past 30 days."

Use-Case Ideas

• Competitive intelligence — compare media coverage between competitors
• PR and media monitoring — understand how your brand or industry is covered
• Trend identification — spot emerging patterns in a specific industry
• Meeting prep — executive news summary with sentiment overview
• Market context — supplement audience or market research with recent news coverage
• Sentiment tracking — understand media tone around a topic or brand

Limitations

• News articles only — no social media posts, academic papers, or proprietary reports
• Cannot access paywalled content
• Single point-in-time analysis, not continuous monitoring
• Cannot verify article accuracy or fact-check claims
• Requires minimum 8 articles to produce a report — very niche topics may not have enough coverage

Input/Output Spec

Input:

• topics (required): One topic per line, supports single, comparative ("X vs Y"), or multi-entity (3-1000 chars)
• cutoff_date (required): Search from this date backwards for a period specified in query (YYYY-MM-DD, must be today or earlier)
• prompt (optional): Additional instructions — focus areas, exclusions, formatting preferences (max 500 chars)

Output:

• Structured news analysis report with:
• Overall media sentiment label and rationale
• Executive summary with key findings
• Per-article breakdowns: headline, source, date, sentiment, summary, key details, significance, link
• Coverage summary (total articles, unique sources, time period, entity balance for comparatives)
• Key themes identified across coverage

Disambiguation Signals

• This agent is for news research and media coverage analysis — it searches, filters, and analyzes recent news articles with sentiment classification
• Route to GWI Spark when: the query asks about consumer attitudes, demographics, or behavior (e.g., "What do millennials think about EVs?")
• Route to Statista when: the query asks for market statistics, revenue data, or industry forecasts (e.g., "What is the global EV market size?")
• Route to Advanced Web Research when: the query needs broader research beyond news — strategy documents, technical analysis, or industry reports (e.g., "How does Tesla's manufacturing process compare to traditional automakers?")
• Route to Company Researcher when: the query asks for a general company overview, not recent news coverage (e.g., "Tell me about Tesla as a company")