Victor Alzate, Chief Product Officer at Ink, shares how Ink is reorganising product development around AI, why its engineers are moving closer to airport operations, and what this shift could mean for airlines.
AI is changing more than how quickly software can be written. It is changing who builds it, how airlines work with technology providers, and which systems they choose to build themselves.
Are we seeing a fundamental shift in how airlines approach AI, or is the industry still experimenting?
We're seeing a shift beyond experimentation. The real shift isn't that people use AI tools; almost every engineer does now. It's that AI is changing how technology teams are organised.
Writing code is no longer what slows teams down. Understanding the problem is. As coding gets faster, direct contact with operational teams becomes more valuable. This brings product and engineering into closer alignment. The people who can talk to an operations manager in the morning and deploy a functional update in the afternoon become the most valuable people in the company.
Airlines are cautious, and they should be. Some major carriers we're already working with are reorganising their internal development around smaller, more autonomous teams. Within a few years, the question won't be "do you use AI?" but "how close are your engineers to the operation?”
How is AI changing who builds products at Ink?
We're moving towards a model where our engineers are product engineers. They don't just work on a single task. They own a problem from the first customer conversation through to production.
AI coding agents make this way of working possible. A conversation with an airline can become a working prototype in hours rather than weeks.
But the important part is what happens next. The same team that built the prototype takes it to production, with the same engineering standards as always: code review, automated testing, security checks, and evaluation of any AI behaviour. Speed on its own isn't the goal. AI tends to amplify whatever discipline a team already has, good or bad.
So AI has changed two things: how quickly we find out whether we're building the right thing, and how few handoffs there are before we deliver it.
Aviation depends heavily on legacy systems. Can AI really help accelerate modernisation?
This is where I've seen some of the biggest practical gains.
Part of Ink's technology is more than 20 years old. For years, only a few people knew how it worked. Today, AI agents help our engineers read, document and test that code far faster, so more of the team can work on it safely.
Instead of a complete system overhaul, we build new capabilities as modern services alongside the existing platform and move functionality across step by step, while the current system keeps running.
The key is that this needs engineers who understand both the code and airport operations. AI can explain what a function does; it can't tell you why a particular station handles a transfer differently. Product engineers who combine domain knowledge with AI-assisted development are what make incremental modernisation realistic.
If AI makes software development faster and more accessible, will airlines start building more of their own technology?
They will, especially larger carriers. We already see airlines building their own Order Management Systems as part of the transformation to Modern Airline Retailing.
However, being able to build something doesn't mean you should. Take sending Advanced Passenger Information to border authorities. It has strict regulatory, security and connectivity requirements, and mistakes are costly. That expertise takes years to build and has to be maintained continuously.
So we expect more airlines, especially larger carriers, to build and own selected core systems and add specialised components from technology providers. The hard part is the integration, and that's where the provider model has to change.
AI is changing the economics of software development. The question for airlines is no longer just what they can build, but what they should build themselves.
How are team structures changing?
Solving the integration challenge is leading to new team structures. The role of the forward-deployed engineer is growing increasingly. It is a product engineer who works directly inside the customer's environment. At Ink, we're building this capability so our engineers can work alongside the airline's own teams: at the airport, with ground staff, and with the airline's IT department.
Their job is to make our components work in production within the stack the airline is building. They don't hand over documentation and leave. They understand the operational problem firsthand, integrate, adapt, and deliver, using AI to move quickly from what they see to working software.
That matters because in aviation, you often only see the real complexity on site: a station-specific exception, a workaround gate agents rely on, or an undocumented integration.
If airlines build more themselves, what should a technology provider bring to the table besides software?
Expertise that's hard to rebuild, and people who can apply it within the airline's own operation.
In the old model, every airline had to work around one large system. The new model is modular products, delivered by engineers who sit close to the operation.
The critical point is what those engineers bring back. Every deployment should make the product better for the next airline, through reusable integrations, configurations and lessons learned. That's the difference between a product company with deployed engineers and a consultancy.
That's the most interesting change: the people who understand the operational problem and the people who build the solution are increasingly the same people.
With AI becoming so embedded in development, are we creating new dependencies and risks?
Yes, and I'd put them into three groups.
First, dependency. We've become reliant on AI remarkably fast. What happens if model costs rise sharply or a provider changes its terms? We design so we can switch models and keep critical knowledge in our own hands.
Second, quality. AI makes it easy to produce a lot of code quickly, and industry research shows that speed without discipline leads to more instability. In aviation, that isn't acceptable. Every change has an accountable engineer, goes through review and testing, and AI-driven features are evaluated before and after release.
Third, data governance. Giving people AI tools is easy; making sure sensitive passenger and airline data is handled correctly is the hard part. This matters even more when our engineers work inside customer environments.
These safeguards help teams put AI into everyday use safely and with confidence.
Beyond software development, where do we see AI making the biggest difference for airlines?
Beyond helping us build software, AI can also change how airline staff use it. That could change the relationship between staff, passengers and their systems.
Today, agents spend much of their time navigating screens, re-entering information and switching between systems. That takes their attention away from the passenger in front of them, often at the worst moment: a disrupted journey or a missed connection.
We want AI to take on that operational complexity, so staff can focus on helping people.
And this connects to how we build. You only discover where that friction really is by being there. When our engineers spend time at the counter and the gate, they see that one rebooking can take twenty clicks. That's why I believe the best airline products of the next few years will be built by engineers who are close to the operation, using AI to turn what they see into working software quickly and safely.
AI doesn't just change how fast we build. It changes who builds: engineers who understand the operation and can take a problem all the way to production.
