TasteLoop
A conversational taste agent that learns what you actually like through natural dialogue, then orchestrates personalized picks across streaming, shopping, dining, and travel — so the platforms stop recommending things you'll ignore.
Adults tired of algorithmic recommendations missing the mark across Netflix, Amazon, Yelp, and Airbnb
- Natural conversation that builds a deep, evolving taste profile (moods, contexts, dealbreakers)
- Cross-platform pick suggestions — asks once, surfaces everywhere you browse
- Taste journal showing how your preferences have shifted over months
- Privacy-first mode: profile stays on-device, only zero-knowledge queries leave the phone
LLM planners can now pick the right personalization tool per user instead of one-size-fits-all ranking, finally making taste-aware recommendations feasible for consumer apps outside the walled gardens of Amazon and Netflix.
McKinsey flags 'agentic commerce' as a seismic shift and multiple AI shopping assistant listicles exist (Shop.app, Buysmart.ai, Rufus, Perplexity Pro Buy), but Consumer Reports 2024 survey of 2,022 US adults shows meaningful discomfort with algorithmic decision-making, which tempers enthusiasm.The agentic commerce opportunity: How AI agents are ushering in a new era ↗5 AI Shopping Assistant Tools To Help You Shop Wisely ↗Consumer Reports survey: Many Americans concerned about AI, algorithms ↗
Cross-platform 'taste profile' agent returned zero direct hits, but the shopping vertical alone is crowded with Amazon Rufus, Perplexity Pro Buy, Google Conversational Commerce on Vertex AI, and Alexa for Shopping — walled-garden incumbents with native data advantages.Amazon Rufus: Amazon's AI shopping assistant gets smarter and more personal ↗Shop like a Pro: Perplexity's new AI-powered shopping assistant ↗Introducing Conversational Commerce agent on Vertex AI ↗
Adjacent AI assistants command $20/mo (Perplexity Pro) and Google offers Gemini Pro/Ultra tiers, demonstrating willingness to pay for AI assistance; however, no evidence yet of consumer willingness to pay specifically for a taste-graph that doesn't own transactional inventory — affiliate/commission economics likely thin.Shop like a Pro: Perplexity's new AI-powered shopping assistant ↗Google AI Pro & Ultra — get access to Gemini 3.1 Pro & more ↗
Agentic commerce and LLM-planner-based personalization are strongly trending per McKinsey and Gartner (Magic Quadrant for Conversational AI Platforms, Aug 2025), but durability is undermined by structural reliance on walled gardens (Netflix, Amazon) that control the canonical taste graphs and could lock out third-party agents or replicate the feature.Gartner Magic Quadrant for Conversational AI Platforms ↗The agentic commerce opportunity ↗
LLM planners + tool-use APIs make the dialogue layer buildable today, but no public APIs expose Netflix, Amazon, Yelp, or Airbnb taste data — the product would depend on scraping, OAuth-scoped read-only access, or affiliate redirects, creating a cold-start data problem and ToS risk.How consumers respond to service failures caused by algorithmic decisions ↗