In an exclusive interview with Harry McNeil, Rob Bassett Cross, a former British Army officer, reveals how his experience in the military and subsequent involvement in investment banking inspired him to found Adarga, an AI software company.

With a mission to empower military organisations, Adarga’s technology transforms unstructured data into actionable insights. Bassett Cross explains how their AI-driven tools optimise information processing, improve efficiency, and enhance decision-making for combat conditions.

By aligning with the priorities outlined in the Defence Command Paper, Adarga delivers AI solutions and enables the UK’s soldiers to maintain their strategic advantage in the ever-evolving landscape of modern warfare.

Rob Bassett Cross (RBC): I served in the British Army for the best part of a decade. I was serving around the time immediately pre-9/11 and post-9/11. Most of my experience comes from Iraq, Afghanistan, and elsewhere worldwide.

By about 2004, we were missing things in the data we had access to. The data we had access to was high-quality intelligence data. It was from all the sensors, the drones, manned-unmanned, and more conventional sources of things like human intelligence.

Long story short, we just weren’t able to analyse it quickly enough or in sufficient depth and detail to spot what I called at the time the needles in the haystack, so things got left on the cutting room floor.

And because we were prosecuting such a high intensity of operational tempo at the time, those things that got lost on the cutting room floor had huge consequences – such as missing targets. 

Weeks later, intelligence analysts would then back solve the situation and we realised we could have done something that night, we could have made a different decision, we could have made a better decision.

Having tried to apply technology back then, which was clunky, slow, and maybe not as cutting edge as we are today – my eyes were opened when I entered the investment banking world for JP Morgan. The technology we had access to was amazing, but I knew it had huge applicability for the targeting cycle or the Intelligence Cycle back in the military.

The motivation then was, how do I take what is now even more exciting technology with things like large language models, natural language processing, and graph technology? How do I take those things that are now possible technically through our data scientists and software engineers? How do I build software for the military and financial services customers where I can help them make better decisions?

That’s our mission. It’s all about enabling organisations to harness greater volumes of information, from more complex sources, in multiple languages. How do I help someone like me back in the day make a better decision and where it’s quantifiable, not just shooting from the hip.

We’re now supplying to the British Army, across the top of the MoD, we’ve worked with the Air Force and Navy. We have contracts elsewhere across government, as well as commercial customers.

We would probably not have had some of those issues and challenges that I had back in Iraq and Afghanistan had we had this technology. We’re just helping to power what the military now call decision advantage or information advantage, accelerating some of those intelligence and analytical use cases.

First and foremost, we’re a software company based in London, built with aspirations to expand to supply that technology to our Allies, firstly in the US and Australia, but then across other allied countries.

Harry McNeil (HM): Thank you for joining us and for introducing Adarga, Mr Bassett Cross.

HM: Adarga specialises in transforming unstructured data into insights. In the context of defence, how is your software helping organisations to identify threat signals and opportunities buried within volumes of information?

RBC: The information environment is becoming a world of mores. There’s much more information. The vast majority of it is unstructured. It’s internal and external. It’s all getting more confusing and more challenging. This is the case in any military organisation.

Getting information into a shape where it’s structured, where a machine can make sense of it, will cut down on the number of human hours spent processing it. We can identify things within that data that would require human time-intensive work.

A good example would be resolving an entity. So, if you were to go to the pub tonight and said I met this guy called Rob, who was talking about AI, and the person at the bar said, what you mean, Rob Bassett Cross. What they’ve performed is a very complex disambiguation computation that we’re good at as humans.

Imagine doing that over millions of sources in hundreds of languages in a split second. That’s what natural language processing, network science, and machine learning can do. What you’re doing is enhancing the ability of that human analyst, making them much better at their job.

The ability to identify events, the ability to help them write their reports, the ability to summarise or synthesise many disparate sources into one single summary, the ability to track live unfolding events, all of that is within the automation of the AI that we put in our software platform.

We train our language models to understand the language of defence and national security. That’s where we apply our in-house specific expertise. We’ve got our team of geopolitical analysts to try to match the workflows. We’re making sure this fits into that human serving in uniform or government, fits into their workflows and helps them do their job better, and also that they trust it. They’ve got to trust the outputs this produces.

HM: Can you elaborate on how Adarga has worked closely with the British Armed Forces to drive AI readiness? Could you elaborate on how AI-driven information tools enhance decision-making in combat conditions?

RBC: There are two sides to AI within the Army. Number one is improving efficiencies, which sounds boring, but it’s important because if we can increase efficiencies for the Army, we can get more individuals back to the frontline jobs.

Then AI has important spearhead capabilities, which include command and control, intelligence decision-making, and enhancing the actual job of the battlefield, but both are important. We help the British Army with accelerating research and analysis, which is quite human-driven.

Analysts spend hours getting their emails, getting all the reports out, and collating information. There’s something called the 80:20 split, where currently, an analyst will spend 80% of their time getting ready all the information and 20% of that time analysing.

We flip that on its head, and that’s the quantifiable time advantage, efficiency, advantage, and accuracy advantage that Adarga delivers. One of the first contracts we won with the British Army was supporting them with their lesson work.

Imagine, if you will, there’s a 20 or 30-year, if not greater, repository of all of the reports that have ever been written about previous operations and training that’s just impenetrable to current commanders, and all of the value and all of the lessons that have been learned from previous conflict and training can now be instantly accessible with the machine doing that heavy lifting of the crunching and analysis.

Suppose I’m planning something in a particular country, or a particular type of operation, or a particular thematic. In that case, I can go straight to this repository and get all the pertinent, relevant responses back and not have to go through the inefficiency of relearning all those valuable lessons.

HM: Thank you for that. I will talk about how the Defence Command Paper Refresh emphasised the strategic importance of adopting artificial intelligence.

HM: How does Adarga’s artificial intelligence software align with the priorities outlined in the paper, and how is the company working to support the UK Defence in this endeavour?

RBC: The Defence Command Paper very neatly identified and was clear on the strategic importance of artificial intelligence. Artificial Intelligence is a foundational technology, and the ability to adopt this and get it working in the real world is a priority.

How do I make a decision? How do I act ever so more quickly against my adversary? For the most part, that’s making sure I can process, understand, and make sense of all the information I have access to, which is bewildering and our platform helps them do that. 

It helps on the battlefield in the here and now. We’re not only supplying our software to the Ministry of Defence and the Army, but we’re also helping them build those foundations that enable them to be good AI adopters and be AI-ready. We’re the partner that produces the tech.

We’ve recently had secondees from the Army and, as you would imagine, we work closely with the Army. We also want to ensure that the skills, understanding, expertise, culture, and systems are all in place so they can adopt this plug-and-play and get on with it.

HM: The commitment to supporting veterans is evident through your Gold Award in the Defence Employer Recognition Scheme. How does your company’s dedication to hiring veterans contribute to its success in the AI sector?

RBC: What those veterans provide is their invaluable military defence or national security expertise. We want to learn from that experience to ensure these models work for those like them who are still there.

We can equip these veterans with new skills for the digital economy that helps the UK drive prosperity, productivity, efficiency, and all those wonderful things.

HM: The use of AI in the military realm calls for a balanced partnership between humans and machines. How does Adarga’s approach ensure that human expertise is integrated with AI capabilities to achieve optimal outcomes?

RBC: We have human experts involved in training these models, so that’s level one. That’s super important. You have to have people who know what they’re doing, have come from that world, really making sure that these models are correctly designed, built, ingrained, and when they’re in production, the management and maintenance of it.

You also need to make sure that you understand and have respect for the workflows that you’re supporting. What AI cannot do and must not do from an ethical and responsible point of view is just come in and say, there’s the bad guy; don’t worry, I’ve decided for you. That’s just not going to fly from a legal, responsible, and societal position.

We’re trying to build trust between the user, the human, the operator, the expert, and the machine supporting and enhancing their workflow. You’ve got to have this intimate understanding of workflow at all stages.

The human understands why the machine is producing certain outputs, understands how that thing was designed and created and why it’s doing certain things, to have the trust and ability to validate the output that it’s giving because we want them to use the output to speed up their entire decision-making cycle, but also have the trust and quantifiable evidence, things like understanding where the source of that information has come from, why is the machine making that connection?

Why has it identified that event? Why is it asking me to read this certain thing now? That needs to be a seamless synergistic relationship. Otherwise, the humans are just not going to trust it.

HM: The last question.

HM: Finally, the stakes are high, and the British Armed Forces opportunity to lead in harnessing AI’s potential is clear. Could you share some essential steps for the UK’s soldiers to maintain their advantage in this evolving landscape?

RBC: The UK already understands the digital infrastructure it needs to adopt AI across the government. What I am optimistic about over the course of the coming months is that the British Army and across the military will become positive adopters of this across all tenants of the business, not only the back office but also the front line, and do that in such a way that talks to some of the human machine teaming aspects that I have touched upon.

There’s no corner of the military this isn’t going to affect. We’re supporting the warfighters, the operators, the decision-makers, the commanders, and that intimate relationship is something that I’m extremely positive about in terms of the door opening wider and wider.

HM: Brilliant, thank you very much, Rob. That was insightful, and it was a pleasure to interview you. I appreciate your time.