Practical AI
for complex
business
We research, build and integrate intelligent systems that run inside real businesses: forecasting models, computer vision, and agents that act under human control.
What we do
Six domains, one method. Select a card to see the project behind it — research, prototype or shipped product.
Computer Vision
Models that read people, motion and events in continuous video, and explain what they see.
View projectFinancial Intelligence
Forecasting models built on market history. Our dairy system reads GDT auction data and prices the next auction.
View projectMolecular & Spectral AI
Inferring chemical properties and candidate structures from spectra: what the instrument records but can't interpret.
View projectMedical AI
Finding early signs of disease in pathology and medical images, with annotation by practising clinicians.
View projectTrusted Infrastructure
A working prototype for digital assets: issuance, lifecycle and permissions, designed to be audited.
View projectEnterprise Integration & Agents
Wiring AI into the systems you already run — integration first, then agents that act, always with a human signing off.
View projectComputer vision for human movement: exercise form, rehabilitation and multi-person activity, tracked in ordinary video rather than a lab. The system explains its feedback instead of just scoring. A seed-stage B2B client now uses it in their own product.
A dairy client provides historical GDT (Global Dairy Trade) auction data; our models forecast the likely price range for each product at the next auction and recommend positions to hold. The same approach carries to asset management and other forecasting problems. The client sees the reasoning behind each forecast, and owns the models.
Research into how machine learning can infer chemical properties and candidate molecular structures from spectral observations.
Detecting anomalies and early signs of disease in pathology images, on an annotation platform built with specialists. Any clinical use requires local data partnerships and independent validation.
An experimental prototype for asset digitisation: lifecycle, permissions and audit trails, designed for regulated use.
View the RWA demo →Connecting AI to the software, data and workflows a business already has, then letting it act: controlled agent workflows, plus applied research on AI for project delivery and organisational risk. Permissions and audit are part of the design, not an afterthought.
Where we stand
Clear about what we are — and what we are not.
The direction
Practical today, agentic tomorrow.
We're building toward that in steps — each one useful on its own, each one still under audit and human control.
Tell us what you’re trying to solve. We’ll give you a straight answer on whether we can help.
Start a conversation →