The job was never writing code. It was solving the problem.

Ask most people what a software architect does all day and they picture someone hunched over a keyboard, writing line after line of code. Julian Cole, Airship’s Lead Architect for fifteen years, barely writes any.

“I don’t really write code anymore,”

That is not because he has moved away from the real work. It is because the real work was never the typing. It is deciding what should be built, understanding the problem it needs to solve, and creating something operators can depend on.

‍

AI can build it. Someone still has to lead it.

Jules joined Airship fifteen years ago as a marketeer, before teaching himself to code and helping to build the platform from the ground up. A lot has changed since then. The purpose has not. His work has always been about helping hospitality operators get real value from their customer data, and that thread still runs through everything Airship builds today.

What has changed beyond recognition is how it is done. AI has made building faster than ever, but not in the way the headlines promise. Software does not write itself overnight from a single prompt. Good software still takes research, planning and careful architecture, and it still needs to be tested, challenged and understood. AI is fast, and it is endlessly confident.

“AI will happily build you the wrong thing very quickly,”

That is why skilled people matter more now than ever. The machine can produce the code. Someone still has to point it at the right problem and stand behind what it makes.

For Jules, that comes down to one word he returns to again and again: ownership. He is relaxed about who, or what, writes a given line of code. What he is firm about is who stands behind it.

“If it ships under my name, I own it and I understand it.”

AI does not remove responsibility from the person building the product. If anything, the speed at which it can create things makes that responsibility matter more. The code still needs to be right. It needs to be secure, maintainable and built for the people who will rely on it. Producing something quickly means very little if it is the wrong thing.

‍

Start with the data.

So, what is the right thing to build? For Jules, the answer starts with the foundation every hospitality business depends on: its customer data.

For years, operators had to work far too hard to get value from it. Answering a single question often meant requesting a bespoke report and waiting a week or two for someone to produce it. By the time the answer arrived, the opportunity to act on it might already have passed.

Ask Jules what he would change about the way hospitality uses technology and his answer is immediate. Get the data clean, complete and in one place. Everything else follows from that. When the foundations are right, operators no longer need to navigate endless dashboards or wait for somebody else to find the answer. They can simply ask.

“The data needs to exist, be all in one place and shaped the right way.”

‍

From searching for answers to simply asking.

This is where Jules sees the biggest change coming, and it is the part he is most animated about. Operators will reach their data in whichever way feels most natural, through a chat window, through the tools they already use and, increasingly, through their voice.

Picture arriving at a board meeting to find the pack already prepared, the right numbers pulled together and the repetitive work taken care of. Or asking a question about customer behaviour and receiving an answer while there is still time to act on it.

“People just want to turn up to that meeting and know it is all there in front of them, with the right numbers in it.”

Airship is already building for that world, opening the platform so its data can flow into the AI tools teams are adopting now, and letting operators reach what they need without changing the way they work. The interface matters less and less. The quality of the data behind it matters more than ever.

‍

Faster to build. Harder to get right.

There is a popular idea that AI will make established platforms like Airship obsolete, because every business will simply build its own. Jules does not buy it.

AI may make software quicker and cheaper to produce, but nobody is safely reinventing their data infrastructure or security over a weekend. The parts customers rarely see, reliability, security and trust, are the parts that matter most. And when those foundations are built properly, they last.

“Nobody is going to be vibe coding their own CRM or reinventing their data security over a weekend. The hard parts of software are still there.”

He has a favourite piece of proof, and it sits inside the family. Much of Toggle still runs on code Jules first wrote in 2018, holding up comfortably at scale years later. By his own estimate, something in the region of £750 million has since flowed through it. That is what good architecture looks like. Not code that simply works today, but technology that keeps doing its job long after the first version ships.

‍

Judgement is the skill that survives every new tool.

Underneath the tools and the standards, one thing still sets great software apart, and it is the least technical of all: judgement. The skill Jules rates most highly is understanding what people actually want, and having the discipline to build that and nothing more. It keeps a product focused, fast and genuinely useful, which is what operators feel every time they open it.

“Think clearly about the problem before you touch the keyboard.”

Programming languages and tools will keep changing, and more of the code will write itself, but none of that alters the purpose of the job. People do not want to spend more time inside your software; they want it to strip away the friction, give them what they need and hand them back their day.

‍

A standard built to outlast the code.

Jules is now preparing to carry that philosophy somewhere new, moving on to help shape the technology vision for a cause he believes in.

“Pointing everything I have learned at a mission I care about felt like the next right thing.”

He leaves behind far more than the code he wrote. He leaves Airship with a team he helped build and a way of working he helped set: think hard before building, own what you ship, and put the operator first.

Our thanks go with him.

‍

See it for yourself

The shift Jules describes, from waiting on reports to simply asking, is not a future promise. It is what Ask Airship delivers today. An operator asks a question in plain English and gets a real answer in seconds, while it still matters and there is still time to act. The full picture already lives in one place, so the answer is one they can trust and act on straight away.

Discover Ask Airship

‍