Flight Centre Travel Group's managing director Graham Turner stood before shareholders at the company's AGM on Tuesday and dropped a number that should make every service business in Australia sit up: the company's AI Trip Support Agent, deployed in August, has resolved 80% of more than 10,000 customer queries at a cost of 11 cents per customer — versus roughly $40 for a traditional support call.
That is not a pilot. That is not a projection. That is a 99.7% cost reduction running in production at an ASX-listed company with $25.7 billion in annual transaction value.
Why this matters beyond travel
The instinct is to file this under "big company does AI." But the numbers Turner presented are more instructive than that. Flight Centre is not a tech company. It is a 44-year-old travel retailer with thousands of consultants across shopfronts, call centres, and corporate offices. If AI agents can deliver this kind of cost compression here, the implications for any business that handles high volumes of customer enquiries — mortgage broking, insurance, hospitality, trades — are hard to ignore.
The Australian average cost of an inbound phone call in a contact centre is $15.29, according to recent industry benchmarking. Flight Centre's $40 figure is higher, which makes sense — travel support calls are complex, often involving itinerary changes, multi-leg bookings, and supplier coordination that push average handle times well beyond a simple balance enquiry. Even against the more modest industry average, 11 cents represents a cost reduction north of 99%.
The full AI stack Flight Centre is deploying
What Turner revealed at the AGM was not a single chatbot but an interconnected suite of AI tools spanning both the leisure and corporate arms of the business.
Co-Consult is the company's proprietary trip construction tool. Consultants set objectives — a family holiday to Japan with a stopover in Singapore, for example — and AI agents find appropriate products across flights, tours, and accommodation. In the past three months, it has handled 20,000 requests and trip constructions, cutting new-consultant planning time from over 60 minutes to under 20.
Workspace Quote, the corporate quoting tool, has reduced the time to send a customer a trip proposal from 17 minutes to six, with the latest version cutting it further to 3.5 minutes for around 290 FCM and Corporate Traveller customers.
Sam and Mel, the company's digital assistants for FCM and Corporate Traveller respectively, are handling real customer interactions at scale. In the first four months of FY27, Sam alone had 134,000 conversations and saved customers around 9,000 hours. Fully AI-powered booking via Sam is now being tested with select FCM customers.
Behind all of this sits a substantial infrastructure investment. Flight Centre's May Investor Day presentation revealed 2,000+ Claude seats deployed across the group, MCP architecture enabling agents to call supplier APIs directly, and partnerships with Google for AI-powered search across the leisure brands.
The vision: compressing travel distribution
The most forward-looking element of Flight Centre's AI strategy is not the cost savings in support. It is what Chief AI Officer Christopher Steiner described in a recent interview as the compression of travel distribution layers.
"What we are seeing in practice is when you set an agent an objective and you say, 'talk to this supplier and here are the rules that you're going to work through,' it doesn't require all that technical infrastructure that has been there really for the last 25 years," Steiner told Travel Weekly.
The argument is that AI agents can increasingly bypass the layers of middleware, global distribution systems, and aggregator technology that sit between a customer request and a supplier's inventory. That could strip millions from distribution costs over the next 12 to 18 months — and it is a competitive advantage that favours companies with proprietary data and direct supplier relationships.
Steiner, who was appointed to the newly created Chief AI Officer role in July after nearly 24 years at the company, is running what Flight Centre calls its Agentic Accelerator program. The mandate: transform the global leisure operating model into what the company describes as an "AI-powered productivity engine."
Flight Centre has structured its AI approach around a 70/20/10 model: 70% of effort on making existing employees more productive, 20% on reimagining roles, and 10% on building entirely new AI-generated revenue streams. As Steiner put it: "The deployment of AI agents will make our people sharper, faster and better informed."
The business case for every service company
Flight Centre's numbers join a growing body of Australian enterprise AI results. Wesfarmers saved 500,000 hours deploying AI agents across Bunnings, Kmart, and Officeworks. Macquarie Bank saved 130,000 hours in seven months with Gemini. Lendi Group's AI agents lifted broker lodgements from 9 to 14 a month.
But the Flight Centre case is particularly instructive for small and mid-sized businesses because the numbers are so concrete. If you are running a service business that handles, say, 500 support calls a month at an average cost of $15 each, that is $7,500 a month in support costs. An AI agent resolving 80% of those at even $1 per interaction (ten times Flight Centre's cost) drops the bill to $1,900 — a saving of $5,600 per month, or $67,200 per year.
Chief product officer Wendie Lee reinforced the revenue side of the equation in an interview with iTnews: "Our average booking value increases because we can sell more holidays with potentially the same number of advisers. And for us as a business, it can lower the cost of service, but also meets the needs of the customer."
That dual benefit — lower costs and higher revenue per employee — is the business case that makes AI agents compelling beyond the headline savings.
What to watch
Three things will determine whether Flight Centre's AI bet pays off at scale.
First, the resolution quality behind that 80% figure. An 80% resolution rate is strong, but what happens to the 20% that escalate? If unresolved queries are creating frustrated customers or extra work for human agents, the net saving narrows. Turner did not address this at the AGM, and it is the detail that will matter as volumes grow.
Second, fully AI-powered booking. Flight Centre is testing end-to-end AI bookings through Sam with select FCM customers. If that works — if an AI agent can reliably book a multi-leg corporate trip without human intervention — it changes the unit economics of the entire corporate travel management industry.
Third, the distribution compression thesis. Steiner's vision of AI agents bypassing traditional travel middleware is ambitious and, if it plays out, would reshape how travel products are distributed globally. But it requires supplier cooperation, and the incumbents in travel distribution have strong incentives to resist.
For now, the numbers speak clearly. An ASX-listed Australian company is resolving customer support interactions for 11 cents that used to cost $40. Whether you are in travel or not, that is the kind of cost curve every service business needs to understand.
Sources
- AI key to shift growth to profit at FCTG — Travel Daily
- Flight Centre targets $325m FY27 profit as leisure recovery takes hold — middle.news
- Flight Centre Global Leisure deploys AI agent to help build holidays — iTnews
- AI could strip millions from travel distribution costs, Flight Centre's AI chief says — Travel Weekly
- Flight Centre to embrace AI agents, ecommerce consolidation — iTnews
- Australian customer service still leans on phone calls — TelcoNews
- Flight Centre Investor Day Presentation — ASX Announcements
