OpenAI quietly rewired ChatGPT’s search algo in December to push ads - and the forensic proof will change how you optimize for LLMs.
You’ve been gaming the wrong ChatGPT since New Year’s. Developers pouring hours into prompt engineering for visibility just got blindsided: OpenAI started tweaking ChatGPT’s algorithm 46 days before announcing ads on January 16, 2026. Our data shows they were live-testing commercial shifts in December 2025.[1]
What happened? Seer Interactive’s forensic dive into 206k ChatGPT responses (Nov 2025-Jan 2026) uncovered the smoking gun. Visibility jumped 4.9pp to 46.2% in December, ‘advice’ queries surged 31%, and rankings favored commercial intent - all before the ad reveal. By early January, re-ranking was fully deployed. Five statistical signals (t-tests, change-point detection) confirm: this wasn’t random.[1]
Why care as a dev? If you’re building LLM apps, SEO, or agentic workflows relying on ChatGPT search, your pipelines are obsolete. Expect paid results to dominate; track portfolio-wide visibility now or watch organic reach tank. Other platforms (Perplexity, Claude) are likely copying this stealth playbook.[1]
Compared to Google SERPs, this is stealthier - no public crawl data to tip you off. Traditional SEO tools miss it; you need LLM-specific trackers like Scrunch AI. OpenAI’s move signals the end of ‘free’ LLM real estate.[1]
Try this today: Run retroactive analysis on your queries (60-90 days back). Tools like Scrunch or custom scrapers. Watch Claude/Gemini data for similar shifts. Question: Will ‘LLM visibility’ become the new SEA? Start tracking.
Source: Seer Interactive