About Cassi

The Origins of Cassi

Cassi’s roots go back to 2008, when CEO Keith Dear was an intelligence officer on his second tour of Afghanistan, working on the campaign against the Taliban’s leadership. He asked a simple question: how would we actually know if it was working?

The result was Beheading the Hydra, a proto-Cassi study. It worked backwards from the outcome (Taliban and al-Qaeda leaders killed or captured to reduce their organisations’ capability) to ask whether that actually worked. It found the strategy was counter-productive. Published in 2011, it became the most-read article in the Journal of Defence Studies, and still sits in the top 10 today.

Assessing Assessments

Soon after, Keith met our Chief Innovation Officer, John Hetherington. Both had read Philip Tetlock’s work on forecasting and became strong advocates for the approach within UK Defence. If this was known forecasting best practice, and UK Government was already committed to probabilistic methods after the Butler Report and Chilcot Inquiry, why wasn’t it being used?

By 2015, that frustration led Keith and John to publish Assessing Assessments. It pointed out that the UK terrorist threat level had sat at SUBSTANTIAL or SEVERE for ten years since 2006. Based on the UK Defence “Intelligence Yardstick,” that implied a 75% or greater daily probability of an attack, a level that has held ever since. Clearly, attacks haven’t happened on 75% of days since 2006. The problem with forecasts detached from reality, in any domain, is that they lead to one of three things:

  • The forecasts aren’t taken seriously, since they’re seen as meaningless, and get ignored.
  • The forecasts get gamed, to win resources against rival areas.
  • The forecasts are taken seriously, and resources get misallocated as a result, either over-estimating risk (as in the terrorist threat example) or under-estimating it.

Often it’s some combination of all three. And if probabilities aren’t derived from a proper system, they’re too easy to politicise, exaggerated for small “p” organisational reasons, or large “P” Party political ones.

Early Advocates of AI in Defence

Around the same time, Cassi’s future founders were recognising how important big data and AI were becoming to decision-making. Keith and our CTO Al Brown were two of the earliest advocates in UK Defence for AI’s central role in decision-making. Al authored the Ministry of Defence’s Human & Machine Teaming Concept Note in 2018. That same year, Keith guest-edited RUSI’s Special Edition on AI, contributing two articles: one on how AI drives rigour into decision-making, another examining Vladimir Putin’s claim that “whoever leads in AI will rule the world.”

More articles and talks followed, and Keith and Al both won fellowships to study AI and decision-making at Oxford: Keith as an experimental psychologist, Al as an engineer. Each kept seeing how difficult and troubled human decision-making was in practice. Keith spent 12 months as Expert Adviser to the Prime Minister, then moved into business, studying for an EMBA at Cambridge and working across Fujitsu’s global corporate structure. Al and John took on more senior defence roles. Across all three, the determination to bring better technology and more rigorous methods to decision-making kept deepening.

Fujitsu C-CAT

Reunited at Fujitsu’s Centre for Cognitive and Advanced Technologies (C-CAT), Keith, Al and John worked with a strong team to push the frontier of “cognitive technologies,” building AI that could draw insight and foresight from data faster and more effectively than humans, while helping bring the UK, Japan, and other allied nations together on the challenge. That work led to a win: the team’s AI, “Cassie,” took first place in the Metaculus Forecasting Cup (Q3 2024), finishing as far ahead of second place as second was ahead of sixth. Cassie was 75% accurate against the human median’s 50%, meaning it outperformed human forecasters by roughly half across more than 300 questions over three months.

As C-CAT’s R&D programmes were mainstreamed and Fujitsu restructured, the team left by mutual agreement to go full-time with Cassi in March 2025.

Creating Cassi

Having seen decision-making up close, in military operations, at the highest levels of UK Government, in international diplomacy, and across global business, Cassi’s founders were determined to take what they’d learned over a lifetime and make it available to everyone. Cassi, and superstrategy, were born.

Cassi was formed to focus on one thing: the continuous improvement of human decision-making, so organisations can navigate change rather than be overwhelmed by it.