At Globus, we’re taking a deliberate approach to AI. Practically, we’ve shifted to capability-driven delivery: we build capabilities once and reuse them everywhere. This may seem counterintuitive in a marketplace where it’s believed that competitive advantage can only be achieved by being the first to deploy agentic AI. But our approach is paying off for us.
Every CIO I talk to is feeling the same pressure to apply AI and apply it fast. Boards, customers, competitors are all pushing us to implement the same strategy. I also feel that pressure. But the past few years have taught me that building the right foundation reduces the risk of moving fast without thinking things through.
If you haven’t heard of us, Globus is an international travel and tour operator. We’ve been in business for almost 100 years, running distinct itineraries across more than 80 countries on every continent for over 500,000 travelers every year.
Although we’re a travel company, two factors are pushing us toward AI that apply well beyond our industry:
Relying on manual processes to address these factors is a liability to our business growth. More importantly, it impacts our mission, which is to turn every journey into a memorable, meaningful experience.
So we started looking at agentic AI as a way to meet our customers’ expectations while solving operational friction to empower our people to drive the largest value.
But we’ve been very deliberate about our approach. Our goal is to do the right things first.
For us, that starts with listening to our employees and customers to understand where AI can solve real friction and create meaningful value. AI is not something we can simply bolt onto the business because the technology is available. It represents a more fundamental shift in how work gets done, how decisions are made, and how the operating model needs to evolve. Like other major shifts in business history, the value will come from being focused and intentional, not from moving fast without changing the way the organization works.
The right strategy starts below the surface: Why data is your secret weapon
By now, we’re all well aware of the cost, both financial and security, of AI sprawl, uncontrolled token consumption, and decisions made without the right human oversight.
Avoiding that starts before you even select an agentic AI development tool. The key to a successful agentic strategy starts with the right architecture, the right data, and the right use cases.
We learned this the hard way.
Early in our modernization journey, we launched an incubator to redesign our pre-trip and on-trip guest apps. As we started connecting those experiences, we discovered our data wasn’t as connected as our vision required. Customer data lived in multiple places, often duplicated, and product data wasn’t consistent. That made it hard to deliver the real-time, connected experiences we were aiming for, and it would have made it impossible to layer AI on top with any confidence.
So we started by investing in Master Data Management (MDM) because the reality is simple:
You can’t scale experiences, and you definitely can’t scale AI, without getting the data right first.
We used OutSystems to set up what we call our unified intelligence platform where we extended our core systems and connected them. This enabled us to keep our heritage while unlocking its value immediately.
Practically, we shifted from a system-centric approach to capability-driven delivery: we build capabilities once and reuse them everywhere.
OutSystems was the ideal platform for this, not just because it let us build reusable patterns (plenty of platforms can do that), but because it’s an agentic systems platform. This means it allows us to build agentic capabilities and reuse them to create systems that work across customer experiences, agent tools, and operations. Everything is grounded in our enterprise context with a single trusted place for connecting and governing the data AI depends on.
With the right foundations in place, the next question was where to start. At Globus, we built a simple filter to decide where AI actually makes sense. We ask four questions:
If an idea clears all four, we move forward. If it doesn’t, it goes on the backlog for later.
One of the Globus brands is Avalon, a fleet of river cruise ships. A common scenario is when water levels on a river make it unsafe for a ship to navigate, and we have to transfer passengers to another vessel, up or downriver. That means updating cabin assignments, inventory, and a long list of other logistics that today are done manually and carry a high risk of error. It’s a perfect candidate for AI because it can surface the right options, highlight the impacts, and provide the information that helps teams move faster.
We’re also building an AI agent that helps travel advisors sell, quote, and service faster. Today that process requires agents to move across several disconnected screens just to answer a customer question. This agent can use customer history and preferences to generate quotes, upsell opportunities, and answers to complex questions. Experienced advisors are then free to close high value sales of trips that often cost thousands of dollars.
In sum, if you’re a technology leader with the same mandate that I do, here are my recommendations before your next AI initiative kicks off:
We haven’t figured it all out yet at Globus. We keep building the foundation and learning as we go. But we’ve learned enough to know which organizations will get value from agentic AI. They’re the ones who built the right foundation and were clear about what they wanted to achieve before they moved at all.
