How Destination Toronto and Brand USA already run on AI

Destination Toronto stopped reporting website traffic as a primary metric. That is not a stunt. It is how they run now. While most of the sector argues about whether AI matters, two of the world’s leading destination organisations already rebuilt around it. Here is what they do differently.

Destination Toronto: the website’s job changed

Paula Port, VP of Global Marketing at Destination Toronto, put it plainly. The website’s job now is answering questions. Inspiration lives in social and with influencers. The team uses AI to spot trending topics and cultural moments the destination can attach to, instead of running traffic campaigns at a number that no longer means what it used to.

That is the shift in one decision. Stop measuring the old funnel. Start feeding the new one.

Brand USA: AI department by department

Janette Roush, Chief AI Officer at Brand USA, runs deep research every morning. “Explain the media landscape in the UK over the last twelve months.” A report that used to take a research team weeks lands in thirty minutes. She wired academic heritage-tourism research into the creative brief for the America 250 campaign using NotebookLM, work the team would not have done by hand.

Brand USA is rolling AI out team by team, building tools around each group’s real work, rather than buying one generic subscription for everyone. Both organisations also hold a hard line on AI-generated destination imagery. The real photo of a real place is the one signal of authenticity no machine fakes, and they will not let AI replace it.

The shift the sector has missed: information agents

From the northern summer of 2026, Google is rolling out information agents. A traveller sets one up to watch the web for them around the clock. Google’s head of Search called it an evolution of Google Alerts. Alerts emailed you links. An information agent maps its own monitoring plan, picks its own sources, and tells you what a change means.

Here is what that does to destination marketing. A traveller planning an August trip sets an agent in June. The agent runs for weeks and builds a picture of your region from whatever it reads, long before the traveller types a question. By the time they decide, their AI already holds a view of your region. Fresh, structured, machine-readable content means that view is built from you. A stale PDF brochure means it is built from TripAdvisor, a booking platform, and a food blogger’s post from 2023.

The moment to influence the decision moved. It used to sit at the query. Now it sits in the weeks before it.

The funnel closed at Google I/O

If you want a date, it is 20 May 2026. Google launched Canvas, a travel-planning workspace that pulls live flights, hotel availability, Maps and local content into one itinerary, and confirmed AI Mode will complete bookings directly with major OTA and hotel partners. TechCrunch summed up the whole announcement in a headline: “Google Search As You Know It Is Over.”

For decades the model ran in a line. Search, browse, land on your site, find an operator, book. The agentic flow removes the ladder. The traveller asks the AI to plan and book, and it does. No browsing step. No listing page. The only organisations with any pull over that booking are the ones whose knowledge shaped the recommendation before the agent opened its booking screen.

What both of them are really doing

Making the destination queryable. A live, structured, machine-readable source for every member, event, itinerary, and bit of local knowledge the region holds. Members keep it current. The DMO sets the standard. AI reads it with attribution. McKinsey and Skift put 70 per cent of travellers as wanting a trip-planning assistant, so the audience for this is already here.

The skills that made a good DMO are the right skills for the job. Local knowledge. Member relationships. Decades of regional know-how. Pointed at making the region queryable instead of distributing brochures.

What to do about it

You do not need to invent this. Toronto and Brand USA already mapped the moves. We pulled them into a seven-move framework, COMPASS, so any tourism organisation can run the same play and take it to a board.

Read the full playbook: COMPASS: The AI-Ready Playbook for DMOs and Tourism Organisations and the practical detail in our 15 FAQs on AI search and destination discovery.

Common questions