As a Product Marketing Director at Azure, I’ve spent the last few years in the thick of generative AI, from early chatbot experiments to sophisticated agentic systems that now reason across modalities. It’s felt like living inside a tech documentary, with every month adding a new chapter. Today, I’m excited to share what might be one of the most quietly profound shifts yet: NLWeb.
Just announced by Microsoft, NLWeb is an open initiative to make the entire web more conversational—enabling users (and AI agents) to interact with websites via natural language instead of dropdowns and keyword searches. For developers and publishers, NLWeb simplifies how you expose your content to AI by using structured data formats like Schema.org, vector indexes, and LLM tools you probably already use.
What makes this even more exciting is the talent behind it: NLWeb is the latest work of R.V. Guha, the creator of RSS, RDF, and Schema.org, who recently joined Microsoft as CVP and Technical Fellow. That legacy of making the web more meaningful lives on here.
This week, I connected with Tripadvisor, one of the early contributors to NLWeb (and my first employer out of college), to learn how they’re thinking about the future of the open, AI-native web. Their team has long been at the intersection of structured data and travel discovery—and with NLWeb, they’re exploring new ways to help users go from vague questions ("Where should I go this fall with kids?") to tailored, high-quality answers, all from their own content.
With the GitHub repo now open and developer guides rolling out, this is the moment to imagine how your website could become more queryable, useful, and intelligent. Whether you’re building for people or AI agents, NLWeb is a new foundational layer for how the web works.
Here is more from the team at Tripadvisor.
Q1: What inspired Tripadvisor to try NLWeb?
At Tripadvisor, we aspire to make everyone a better traveler by offering contextual insights, guidance and recommendations. As such, we've been actively exploring ways to make travel search more intuitive and flexible. NLWeb offered a compelling opportunity to test how natural language input and conversational interactions could simplify complex queries and reveal more of our inventory without requiring users to master a series of filters or form fields. This makes it easier for travelers to express intent and get relevant insights faster.
Q2: How did the setup process go for your team?
The setup process was straightforward. We quickly integrated the NLWeb API into our frontend UI, and on the backend, we connected it to our existing vector database (Qdrant) with minimal changes. Overall, it was easy to integrate into our existing tech stack which is already optimized for search.
Q3: What’s a query or interaction that made NLWeb really click for Tripadvisor?
A moment that really clicked was when a user typed: “Dinner for Father’s Day in Seattle with good wine near Pike Place.” Traditional search would require manually applying filters for location, occasion, cuisine, and wine quality, often across several menus or screens. With NLWeb, that full intent is captured in one simple, natural sentence, and the system understands and responds accordingly. It enables more expressive, human-centric search and helps us surface relevant results by tapping into our very rich knowledge graph consisting of both structured and unstructured data in ways traditional search can’t. Plus, not having to add, update and maintain hard tags for filters is another appeal for our data and AI teams.
Q4: How are you blending NLWeb with your current design or customer journey?
We’re currently exploring how to integrate NLWeb into our traditional typeahead search experience. The goal is to support longer, more natural queries while enabling users to interact with results in a way that feels more interactive. Since NLWeb works independently of the UI, we can integrate it fairly easily into our existing site design, including potentially into our Trips product or AI Trip Planner.
Q5: If NLWeb reaches its full potential, what could that unlock for Tripadvisor users or the web at large?
We see natural language as a foundational shift in how people interact with travel platforms and potentially the web overall. Our moonshot thinking is that we can and should anticipate, evolve and exceed our traveler expectations, meet them where they are and remove any interaction barriers. If NLWeb and tools like this continue to evolve, they could render many traditional UX patterns obsolete: no more checkboxes for amenities or rigid dropdowns for price. Instead, a user could just say what they want, and the system understands and responds intelligently across form factors, different modalities, and may be even language. That opens the door to radically more intuitive and expressive digital experiences that are customer fist. This is a massive win for everyone - all users and all businesses involved in this ecosystem.
At Tripadvisor, we aspire to make everyone a better traveler by offering contextual insights, guidance and recommendations. As such, we've been actively exploring ways to make travel search more intuitive and flexible. NLWeb offered a compelling opportunity to test how natural language input and conversational interactions could simplify complex queries and reveal more of our inventory without requiring users to master a series of filters or form fields. This makes it easier for travelers to express intent and get relevant insights faster.
Microsoft