CompleteFlow is an agentic AI platform that puts security first.

It processes documents, runs workflows, and answers questions across your organisation.   The platform can be deployed on your own infrastructure with full audit trails and human-in-the-loop controls, connecting to 10,000+ systems via the Model Context Protocol

CompleteFlow lets your people focus on high-value judgement work while the platform handles the rest.

We built AI tools to make sense of terabytes of unstructured government documents dating back to the 1980s.

Using knowledge graphs and custom algorithms, we turned vast archives into clear insights that help shape better policies.

By managing diverse stakeholders and handling complex data formats, we delivered a solution that works at scale and drives real results.

It takes an average of 15 weeks from initial symptoms to diagnosis of cows milk allergy.

We delivered a conversational AI that acts as a companion for parents throughout this journey, helping them to track symptoms and produce a comprehensive report to take to the GP, significantly reducing the average diagnosis time.

Our team focused on optimizing performance for fast responses, integrating the solution with Danone’s tech stack, and ensuring compliance with strict healthcare data privacy regulations.

Over 90% of users would recommend our product to a friend.

What is the impact of COVID-19 on global instability? We have developed data insights tooling and strategy to address this question as well as many others for analysts across multiple government departments.

We successfully delivered operational dashboards that demonstrate the ability to extract actionable insights from unstructured text data, along with a long term strategy for building capacity.

We developed Economy.os - a platform for simulating complex systems using multi-agent systems and reinforcement learning.

This novel approach was applied to modeling tokenomics for blockchain based economies.

Solved for multiple equilibria and coordination problems traditional methods miss, without requiring training data.

‍

‍

Bit Bio raised over $100M to develop  cell coding technology that enables a new generation of cell therapies.

We are helping them to build a data driven ecosystem with radical transparancy that empowers employees to make better decisions and creates a flywheel for growth through compound improvement.

Our work includes organisation-wide data strategy and implementation of core AI infrastructure.

‍

We built Groupseer, a social media intelligence platform that uses unsupervised machine learning to map and segment online communities at scale.

The platform builds and clusters graphs of social relationships, then uses language models to summarise the content and interests of each community segment. This gives social media marketers a data-driven view of their audience that goes far beyond basic demographics.

Our solution uses Databricks, Spark, and Elasticsearch to scale horizontally in the cloud and handle terabytes of data. The core technical approach has since been patented.