Intentful.ai

Webinar Recording and Transcript: What Visitors Really Ask AI

Intentful Insights Team
August 26, 2025
20 min read

This piece adapts a webinar created for Destination Marketing Organizations (DMOs), delivered by Marina Petrova, CEO & Co-Founder of Intentful, on August 26, 2025. The transcript has been edited for smoother reading while keeping the speaker’s intended meaning intact.

Marina Petrova introduced the session by explaining that it would draw from thousands of real visitor questions submitted to AI on destination websites, then translate those patterns into practical guidance for DMOs that want to be easier to find and more useful in an AI-mediated environment.

The session focused on three areas:

AI is reshaping how people search for destination information

AI has moved well beyond early adoption. It is changing the way people engage with content, brands, and places. For DMOs, that shift can feel complex, but it also creates a new opening: serving individual visitors with more personal answers while still operating at scale.

Built from practical AI work

Intentful’s perspective comes from building AI systems in the real world rather than from abstract theory. The team works directly with APIs, crawlers, retrieval behavior, embedding pipelines, and the ways AI systems search for, retrieve, and interpret information. That system-level experience informs Intentful products including AI-search-driven Assistants and Generative Response Ads, where AI needs to locate and deliver brand-aware answers instantly.

AI changes every day. Intentful treats that pace of change as part of its operating model. Since 2021, the team has worked deeply on AI discoverability across multiple LLMs and has continued adapting alongside the technology.

In addition to destinations, Intentful currently works with organizations in travel, performing arts, consumer packaged goods, telecommunications, and agencies, with a growing international presence.

Methodology

The findings in the webinar are based on 15,000 anonymized questions submitted by real users through Intentful technology and one of its products, the Intentful AI assistant embedded on DMO websites.

We never collect any personally identifiable information. We only know the question that was asked, but we don't know who asked it.

Some visitor-question examples on the source page are shown in their original, unpolished form, including typos and grammar mistakes, because that reflects how people naturally use AI. Other examples are slightly adjusted to remove sensitive details while keeping the underlying intent.

The broad pattern: visitors ask nearly everything

At a high level, visitors ask AI about almost any topic connected to a destination. Their questions range from practical needs like parking, opening hours, and events to highly personal concerns, unusual curiosities, accessibility needs, and local-history questions.

The source page includes many anonymized visitor-query examples from travelers, residents, prospective vendors, eventgoers, accessibility-conscious visitors, families, pet owners, food-focused visitors, and people seeking confirmations or support. Across those examples, recurring themes include logistics, cost, timing, access, local recommendations, event details, transportation, personal context, safety, history, and questions that sound like they would normally be asked of a knowledgeable local.

Examples of the kinds of visitor questions include:

For destination teams, the key question is simple: if someone asks this type of question on your website, does the website actually contain the answer? If the answer is missing, neither AI nor the visitor can reliably find it.

Visitors speak to AI like they are speaking to a trusted local

Questions span parking fees, haunted sites, thrift stores, bike races, July 4th drone shows, ADA beach access, events, local history, and more. People do not usually type rigid keyword strings anymore; they ask complete, conversational questions and expect context-rich responses.

Visitors also interact with AI in a human way. They greet it, thank it, joke with it, test it, and sometimes express frustration, even while understanding that it is not a person.

Every query reflects a need for information

Most destination AI questions fit into three broad information needs:

Question categories and examples

Core logistics and practical details

Visitors often ask basic questions that determine whether they can plan confidently. Examples include:

The practical implication: make sure the essentials are easy to find and up to date, including hours, reservations, pricing, parking, accessibility, and seasonal timing.

Direct, transactional requests

Many people use AI with a clear action in mind. Examples include:

The pattern: transactional intent is growing as people realize AI assistants can handle more than old scripted chatbot-style exchanges.

Human curiosity and destination context

Visitors are not only looking for listings. They also want explanation, background, and local context. Examples include:

The takeaway: people want the story, the reason, and the context, not just a directory entry.

Follow-up checks after an action

People often come back with questions after a purchase, submission, booking, or decision that may have happened elsewhere. Examples include:

The implication: content should anticipate follow-up questions, even when the original transaction occurs through another system or organization.

Accessibility and personal context

Visitors frequently provide personal details because they want advice that fits their situation. Examples include:

The content need: include accessible routes, mobility notes, equipment information, terrain difficulty, pet policies, and dietary considerations wherever they matter.

Local questions from residents and partners

Not every person using a destination website is a tourist. Locals may ask:

The takeaway: destination websites also serve residents, businesses, partners, and community stakeholders, so content structure should support those users too.

Events and near-term happenings

Events are often tied to immediate dates or short planning windows. Examples include:

The practical need: event information should stay fresh, structured, and filterable by date range.

Food and drink

Visitors use AI for food recommendations with specific preferences or occasions in mind. Examples include:

The implication: listings alone are not enough. Add attributes that help both people and AI decide, such as kid-friendly, outdoor seating, gluten-free options, and other useful selection criteria.

Lodging, transportation, and parking

Travel-planning questions often mix lodging, movement, and access. Examples include:

The key issue: many users assume the destination site can provide real-time information. Be clear about what is general guidance and what is live or time-sensitive, and point people to authoritative sources where current data is required.

Trip planning does not happen in a straight line

The travel funnel exists: dream → plan → book → visit. But travelers do not behave as if each step is a clean checkpoint.

People plan across multiple sessions, change topics, and return with additional questions. Destination content should be built so any page can serve as an entry point and so answers make sense without requiring the visitor to start elsewhere.

How AI chooses what to answer, in plain language

If something is not published, AI cannot know it. Some systems may try to fill the gap with a guess, while stronger systems may refuse to answer. In both cases, missing information creates missed opportunity.

AI relies on available facts rather than on your assumed authority. It does not automatically understand that your organization is the official DMO unless your content and markup make that clear. At present, AI often favors information it can interpret as clear.

Retrieval is based on context, not only exact keywords. Content is split into chunks, transformed into vector embeddings, and matched according to meaning.

Machine-readable structure matters. Headings, lists, tables, schema, sitemaps, and robots settings all help AI systems access and parse your pages.

AI may not keep trying if it fails to find the answer. Additional retrieval attempts cost compute, so destinations should not assume there will be another opportunity to surface the right information.

Two major failure modes when AI cannot locate the answer

Pages AI cannot read well

This can happen when pages have weak structure, missing or flawed sitemaps, robots issues, or content that is inaccessible to crawlers and retrieval systems.

Facts that are absent or incomplete

A page may be visually beautiful but still unhelpful if it does not include concrete details such as hours, fees, parking, accessibility information, or other essentials.

Actionable recommendations from Intentful

Intentful’s guidance reflects both the system-level view of how AI ingests and organizes content and the human view drawn from 15,000 authentic visitor queries.

Closing

AI gives destinations a way to support each visitor with specific, contextual responses. Make the site understandable to machines, support the destination’s story with detailed facts, and create a reliable process for keeping time-sensitive information accurate.

About this article

Intentful Insights shares perspectives on AI, brand strategy, and the shift from one-way messaging toward two-way customer communication.

This article adapts the What Visitors Really Ask AI webinar presented by Marina Petrova, CEO & Co-Founder of Intentful. It covers real visitor questions from 15,000 anonymized queries on DMO websites, explains how AI retrieval works, and offers practical recommendations for destination marketers.

Intentful is commercially deployed since 2021, working with organizations in travel and tourism, performing arts, CPG, telecommunications, and agencies globally. Contact: [email protected]

Visit the What Visitors Really Ask AI: Webinar Recording and Transcript — Intentful Insights page →