Secure Agentics co-founder and chief executive Max Corbridge tells Emily Moore about a career which has taken him from legally breaking into banks to creating a ‘security monitoring system for AI’
WHEN Max Corbridge told his mother that his week had included a “night shift in which he had broken into a bank”, it is hardly surprising that she was somewhat concerned by his news.
But, for many years, hacking into organisations’ systems and carrying out the occasional 3am physical break-in was all in a day’s work for the former Victoria College head boy.
“It was great fun and, most importantly, it was all legal,” stressed Max with a smile, who became an ethical hacker after a chance conversation with a friend who was working for the Jersey Financial Services Commission.
Indeed, the surfing enthusiast had never considered a career in the technology sector until, after graduating from Exeter University, he realised that his original goal of joining the Army no longer seemed quite so appealing.
“I was in the Combined Cadet Forces at Victoria College and spent most of sixth form convinced that I was going to join the Army as soon as I finished school,” said the co-founder and chief executive of Secure Agentics.
“However, the head of sixth form persuaded me to apply for some university courses, which I did without visiting the places or really looking in any detail at the courses. At the last minute, I was out surfing with a friend’s older brother and he suggested that I should go to university because ‘the Army would always be there’.”
Unclear about what he wanted to do, Max therefore enrolled at Exeter – “because there was a direct flight from the Island” – to study French and Spanish.
“I had a great time, including an amazing year in Catalonia, but when I graduated, I still wasn’t sure what I wanted to do,” he said. “The Army didn’t have the same appeal it had once had, but I couldn’t imagine going into finance or joining a post-graduate scheme. A friend from Vic was working in the cyber team at the JFSC and he mentioned that you ‘could get paid to hack into banks’.”
Immediately intrigued by the idea, Max started “doing some digging” and quickly decided that penetration testing, or ethical hacking, was something he could “get excited about”.
“I spent two weeks shadowing my friend and decided that was the career path for me,” he said. “The only problem was that ethical hacking is a very technical role in a very technical world and beyond having played a few computer games while I was growing up, I knew nothing about computers at all.”
Never one to shy away from a challenge, Max, who was back in Jersey last month to give a fireside talk at Digital Jersey, embarked on his next phase of learning.
“I locked myself in my room, studying for 12 hours a day, seven days a week for five months, teaching myself everything I could about ethical hacking, after which I got a job in a leading London-based ethical hacking consultancy,” he recalled.
But while instantly captivated by what he was learning and recognising that it offered “good long-term career opportunities”, Max admits that not for one moment did he imagine the path that career would take.
“When I started in that role, I did it because it was fun, offered a stable income and gave me the opportunity to work remotely,” he said. “I also knew that it was going to be really hard, and I’ve always been attracted to big challenges and teaching myself new skills.
Although I didn’t know it at the time, taking on something so challenging and getting into a really technical role without a technical background was the best thing I could have done, as it proved that things which don’t initially seem doable can indeed be possible.”
It is that philosophy which gave Max not just the confidence “to change jobs and retrain from scratch every couple of years because he didn’t want to get comfortable” but also to launch his own business last year.
“From my very first day at work, I knew that, at some point, I wanted to do my own thing,” he reflected. “I was lucky in that the founders of the first two companies I joined were still there, so I asked them both what their founding journey had been like.”
Those insights, combined with his experiences as an associate penetration tester, consultant, senior and then head of team, eventually persuaded him that he had “the network and experience needed” to give in his notice and focus on his own business goals.
“I quit my job in March 2025 and then did some contracting for a while, as that enabled me to start a business while building up as much money as possible,” he said. “During that time, I worked from 5.30am every day, carrying out two full-time jobs while starting Secure Agentics, making sure that I had enough money behind me to sustain the business for four months.”
Acknowledging that it was a “demanding time”, Max admits that that was just the first of many “demanding times” ahead.
“As soon as I had launched the business and hired my first engineer, I had to focus on fundraising, as we only had four months of runway before we were dead in the water,” he said candidly.
Fortunately for Max, his vision of a “security monitoring system for AI” appealed to investors and, following that successful first fundraise, he quickly brought his first product to market.
“The best way to describe Secure Agentics is as a security bubble around your AI, protecting you and your interests,” he said. “When you ask AI to do something for you, our systems makes sure that it is doing what you asked it to, isn’t misbehaving and hasn’t been hacked.
“For example, a lot of people ask AI to help them with their personal finances. The problem is that because AI is still in its infancy, we don’t know how to use it securely or to stop it from doing things we didn’t intend. Plus, AI is fundamentally vulnerable to attack, which means that it doesn’t know the difference between your instructions and those given by a malicious actor.
“Therefore, if you ask AI to transfer money to your partner, AI will listen and do that. But if someone hacks into your system and asks AI to transfer money to their account, it will try to do that, as no one has yet found a way to make AI listen only to you.”
While Secure Agentics does not solve that problem, Max says that the open-source product, which is free for individuals, monitors what AI is doing and alerts people to any suspicious activity.
“If you continue the above example, if you are using AI to help you manage your finances, you can set up Secure Agentics with an instruction that says, ‘I want you to make sure that AI only ever sends money to my partner’. Therefore, if an AI agent decides to speak to an attacker and receives a malicious instruction asking it to send money elsewhere, our system will stop AI from taking those rogue instructions.”
Saying that they are the “first people in this space to offer a free, open-sourcing product”, Max adds that the initial response has been “very positive”.
“We had just under 4,000 downloads in the first five weeks,” he said, “and lots of people in the community have offered to build new features for it and help to expand it. For us, as a company, that is like nirvana. That response is what every open-source company is aiming for, but it is incredibly rare.”
And it is a response which has not just benefited the product but has also provided a recruitment opportunity.
“We launched a version of the product that integrated with a programming language called Python. Someone, who are we now looking to hire, has already built a version which integrates with another language,” he said.
Having gone to market initially with a free product, Max and his team are now building a version of Secure Agentics designed for businesses.
“That is how we will generate revenue,” he explained, “but we felt it was important to be a community player in this space and not just look for commercial gain, as everyone else has focused purely on creating products they can sell.
“As well as targeting individuals, we are keen to reach developers and coders who can build our monitoring system into anything they are building with AI.”
Offering the product free of charge is not, says Max, the only way in which Secure Agentics differentiates itself.
“We were the first ones to look at the process which AI goes through when carrying out an instruction,” he said. “Last summer, I saw some research from OpenAI and Google DeepMind which highlighted that when monitoring AI, it’s not just a case of looking at what AI is doing but also what it is thinking.
“When you give AI a task, you can see and read its thought processes and what it is going to do. Going back to the money example, if you ask it to transfer money for you, you can see it going through the motions of how it is going to access the bank account, how it will find out who the recipient it, how to enter the recipient’s details and then make the transfer.
“As soon as I saw that research, I thought how cool it was, and that’s what inspired me to quit my job and essentially reverse engineer the research paper to figure out how they had done it. That’s when I saw the potential for a company and hired my first person.”
Having founded the business with the late Steve Street, who founded MindGard and spent 20 years building and selling cyber-security companies, Max is now deeply entrenched in the next round of fundraising.
“This is a world which moves very fast; it’s an insane market,” said Max with a smile. “But I don’t think you will ever find a more exciting time to start a business. AI is exploding and completely changing how you run a business, but while this is an incredibly exciting, empowering, innovative and life-changing technology, I am also very worried about the fact that the security side of it has been largely ignored so far.
“I spoke on BBC News recently about the risk of biological warfare now that AI can produce gene sequences for viruses which have been extinct for many years. That means that, in theory, someone could revive an extinct virus which no one is vaccinated against.
“Add into that the fact that the companies making AI are obliged to maximise profit and so don’t have the moral resolve to slow down and do things correctly, and the risk increases further. That’s why, whatever happens with Secure Agentics, I was determined to fight the good fight and focus on the security side of AI.”
