A guest asking about wedding venues at 10 pm is a lead. A spike in confused ticketing questions is a UX problem. Most organizations are logging these as support tickets.
Most attractions and destinations still treat AI chat as a support tool. Something that answers hours and ticketing questions so the phone line rings less. That view misses what is actually happening in these conversations. Every question a visitor asks an AI agent is a signal about what they want, when they want it, and what would get them to spend more. Organizations that treat that signal as a customer service log are sitting on data that marketing teams would pay for.
A good example of this comes from Franklin Park Conservatory and Botanical Gardens in Columbus, one of the founding members of the Agentic City program launched by Satisfi Labs this year with Experience Columbus. Over a three-month period, Franklin Park’s AI agent fielded thousands of guest conversations, and the patterns inside those conversations tell a much bigger story than “we answered some questions.”
Start with timing. A large share of the conversations occurred at 10 pm, 11 pm, and later hours, when no staff member is on site to answer the phone. People were asking about hours, ticket availability, membership eligibility, and event registration well after the gates closed. For an attraction, that is a staffing problem solved without adding staff. For a marketer, it is something else entirely. It is a record of intent that would have evaporated by morning.
Also Read: HubSpot’s Aja Frost on Marketing in the Age of AI Search
That intent shows up most clearly around events. When Franklin Park hosted a Gabby’s Dollhouse Meet ‘n’ Greet, the agent absorbed a massive spike in volume, easily a third of all conversations during that window. Guests asked how to register, why a second ticketing step wasn’t appearing, and whether the event was sold out. The agent guided families through a process at scale, without additional hires. But it also surfaced something the events team hadn’t had visibility into before. Teams could see exactly where the ticketing flow was breaking down for real guests, in real time, and in their own words. What began as a customer service interaction ultimately became a source of insight for both the events and marketing teams.
The same pattern revealed something traditional analytics tools rarely catch. Guests asked repeatedly about Museums for All and social services-based discount programs, including questions about specific managed care plans and out-of-state eligibility. These are not casual browsers. They are people actively trying to visit the Conservatory, and the questions show how conversational AI can help visitors navigate access programs and find information that might otherwise be difficult to obtain. That is an audience insight with implications for outreach, programming, and community partnerships that extend well beyond admissions.
Then there is the revenue conversation hiding inside the customer service conversation. Guests asked about weddings, birthday parties, graduation parties, and barn rentals, often with specific dates, venues, and budgets in mind, and frequently at night when the events team was unreachable. Others asked whether the special exhibit changes seasonally, when a specific show is running, and how much the butterfly larvae in the gift shop cost. None of that would show up in a sales report. It shows up only in the conversation itself, and only if someone is positioned to capture it.
Also Read: The Agency-Led E-commerce Model Is Changing
This is the shift marketers need to make. Unlike traditional FAQ tools, AI-powered conversations do more than reduce support volume. They’re giving marketers direct visibility into audience intent. A guest asking about wedding venues at 10 pm is a lead. A guest’s question about whether an exhibit rotates seasonally is a programming signal. A guest asking about a discount program is an equity and outreach signal. A spike in confused questions about a single event is a UX problem marketing can fix before the next campaign.
Every one of these signals already exists inside the conversations attractions are having with visitors right now. Marketers have spent years trying to infer customer intent from clicks, page views, and conversion data. Increasingly, customers are telling us exactly what they want. The organizations that learn to listen will have an advantage that no dashboard can provide.
