Google's AI Mode gets better at understanding visual prompts (www.engadget.com)

Google announced an update to AI Mode that improves how the chatbot handles visual prompts by applying its "query fan-out" technique to images and videos. Instead of returning long blocks of text when you ask about an image, the system now runs multiple parallel searches behind the scenes to better interpret style, composition and intent — for example, understanding what you mean by "moody but maximalist" bedrooms — and surfaces more targeted, visual results. The feature is multimodal, so you can start a conversation with an image or video and immediately get a richer visual response. This change is significant for AI/ML users and product teams because it boosts retrieval relevance and conversational refinement in visual search scenarios, particularly shopping: AI Mode can find items that match nuanced constraints (e.g., “barrel jeans that aren’t too baggy”) and then respond to follow-up filters like “more ankle length.” Technically, the update leverages background fan-out queries to disambiguate user intent and prioritize image-based answers, improving UX for multimodal search workflows. Google is rolling the update gradually, so the new experience may take a few days to reach all users.