{"rewrite":{"id":"r_ba7ac545ebbd52cd33adc861","clusterId":"c_cf2be99d2bc63c775ec1649f","slug":"renue-ai-agent-now-bases-proposals-on-human-verified-win-patterns","model":"deepseek-v4-flash","headline":"Renue AI Agent Now Bases Proposals on Human-Verified Win Patterns","summary":"Renue updated its advertising operation AI agent to include only human-verified win patterns as evidence for proposals. The feature, added to the service launched in February 2026, structures suggestions with the target campaign, expected effect, success criteria, and the verified learning that supports it. The goal is to accumulate institutional knowledge within the client's account rather than training renue's AI.","whyItMatters":"The update addresses the lack of traceable evidence in AI proposals and the loss of institutional knowledge when staff changes, by making human-verified learnings the sole basis for suggestions.","webCardHtml":"\u003cp\u003eThe update structures each proposal with the target campaign, expected effect, success criteria, and the win pattern that supports it. Proposals require human approval before execution, and budget changes are never automated. The service, which uses Anthropic\u0026#39;s Claude Code, supports Instagram, Facebook, Google Search, YouTube, X, and TikTok. Daily improvement proposals are generated from Google Ads and Meta Ads performance data.\u003c/p\u003e\u003cp\u003eRenue designed the system so that learnings stay within the client\u0026#39;s account and are not used to train the company\u0026#39;s AI. The company positions the feature as a way to turn ad spending from an external cost into an internal asset, with a clear audit trail of what was proposed, who approved it, and what was changed.\u003c/p\u003e","blueskyPost":"Renue's update deliberately walls off cross-client learning. The AI agent now only uses human-verified patterns per account, meaning each client's institutional knowledge stays siloed rather than feeding Renue's model.","twitterPost":"Renue's AI agent now builds per-client knowledge from human-verified patterns, not from cross-account training. Institutional memory stays with the client.","threadsPost":"Renue's AI agent update makes a deliberate design choice: it only uses human-verified win patterns per account, and the goal is to accumulate institutional knowledge within the client's account, not to train Renue's own AI. That means each client's learnings stay siloed, not shared across the platform.","newsletterBlurb":"Renue added a feature to its advertising operation AI agent that uses only human-verified win patterns as evidence for proposals. The update structures suggestions with expected effect and success criteria, and requires human approval before execution. Learnings accumulate within the client's account and are not used to train renue's AI.","attributionJson":"[{\"source\":\"ASCII.jp\",\"url\":\"https://ascii.jp/elem/000/004/419/4419815/?rss\",\"title\":\"広告の勝ちパターンをAIが分析し、会社ごとの改善を提案する機能を「広告運用AIエージェント」に追加\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":5870,"outputTokens":1516,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1784098775,"createdAt":"2026-07-15T06:53:22.000Z","publishedAt":"2026-07-15T06:57:28.000Z","updatedAt":"2026-07-15T06:53:22.000Z"},"cluster":{"id":"c_cf2be99d2bc63c775ec1649f","canonicalTitle":"広告の勝ちパターンをAIが分析し、会社ごとの改善を提案する機能を「広告運用AIエージェント」に追加","representativeArticleId":"a_f259cc9a2ee1786e76dafde3","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[],\"studios\":[],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-07-15T03:00:57.000Z","lastSeenAt":"2026-07-15T03:00:57.000Z","updatedAt":"2026-07-15T06:57:29.000Z"},"attribution":[{"source":"ASCII.jp","url":"https://ascii.jp/elem/000/004/419/4419815/?rss","title":"広告の勝ちパターンをAIが分析し、会社ごとの改善を提案する機能を「広告運用AIエージェント」に追加"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":[],"studios":[],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":null}
