The Astut Story So Far

As well as building great software the story of Astut is about realizing that modern artificial intelligence was missing something fundamental and setting out to address what is possible?

For many years, the AI world has been focussed on optimization: finding the “best” path, the “most likely” answer, or the fastest route. All based on utilising large sets of data with the aim to get to, or achieve, a pre-defined goal.

Novelty Search

Back in 2015, decades of research specialising in complex systems and optimisation led Pete Grindrod, professor in the Institute of Mathematics at the University of Oxford, to become interested in novelty search. He was considering how to systematically explore enormous spaces of possibilities when neither the geometry of the space nor the location of valuable discoveries is known in advance.

The pioneers of novelty search, Kenneth O. Stanley and Joel Lehman assert that many significant discoveries emerge through open-ended exploration rather than direct optimisation. In fact, knowing where you want to get to can even hinder you from getting there.

By extending novelty search far beyond evolutionary algorithms, Pete came to realize the principles of discovery didn’t require complex, high-level mathematical theories; they required a way to push exploration into unknown territory. Rather than viewing novelty search as a technique on its own, he began to regard it as a general mathematical framework for exploring the spaces of greatest practical importance: “the set of all possible answers or responses to a specific question”. At the outset, such spaces are unknown and cannot be parametrised or explicitly defined. This represented a new challenge for novelty search methods: how to explore, sample, and map. However, one usually does know what cannot be: the answers to any specific question are all subject to a wide range of logical constraints: so two complementary subfields of mathematics must work in tandem.

When the stakes are high

Leaders and boards are increasingly expected to make decisions in the face of unprecedented events. When a company enters a brand-new market or deals with a geopolitical crisis, there is no historical data to rely on and no single “correct” answer to calculate. In these moments, standard AI is useless because it tries to find a “winner” among existing options, but what these leaders actually need are options and a map of what is possible.

By utilising novelty search along with inverse reasoning Pete envisioned a system that didn’t just guess what might happen next, but one that could systematically map out a vast, diverse range of credible strategic options. He began to call this Explorative AI.

From Theory to Product

In early 2025, Pete decided to take this vision beyond the towers of Oxford University to become the CEO of a new spin-out company and Astut Ltd was born. By the summer of 2025, the company secured immediate seed funding led by Sure Valley Ventures, East X Ventures, and UKI2S.

Yet a visionary idea is only half the battle. To turn these concepts into a practical, usable platform, Pete needed a team that could bridge the gap between academic theory and real-world engineering. This is where the founding team of Sharon Whiteside, Chris Wilson, and Danny Rayan came in, joined later by Katie Ross and Jack Hindmarch.

In September 2026 they started work to produce prototypes from the theoretical concepts. They weren’t just coding; they were collaborating on a process of invention, solving the massive challenge of how to build a software platform that could discover, hold and organize thousands of strategic ideas simultaneously. The result is the Astut Hypothesis Engine® launched in June 2026.

Vision for Growth

From the offset Astut engaged with key stakeholders working on high-stakes problems in fusion energy, finance, and cyber security to prove that Astut Explorative AI® works in a vast and varied set of commercial sectors.

The ambition quickly evolved. Astut didn’t want to be a traditional consultancy. They adopted a channel-first strategy, with the aim to become the “Dolby” of strategic AI. Just as customers buy devices “Powered by Dolby,” Astut aims for their Explorative AI technology to sit inside the solutions of major systems integrators and advisory firms worldwide. To this aim they were delighted to announce their first channel partner: Soma Technologies in July 2026.

The Future

As of September 2026, as Astut reaches its first birthday, it has established a clear vision. Explorative AI is not just a tool. It is a new category of AI that stands alongside prediction, optimization, generation, and agency.

The story of Astut is still being written, but the mission remains consistent: to provide leaders with the clarity they need to navigate the twenty-first century, not by replacing their intuition, but by giving them a clearer, broader view of the world of possibilities.