AI Innovation Lab

Core Purpose

The AI Innovation Lab is fundamental to Decision Lab’s success. Its mission is to explore what’s next in AI, simulation, and optimisation, and translate those possibilities into tangible value for our clients and our business.

Our pipeline of next-generation capabilities provides our teams with a unique advantage, allowing them to deliver more advanced, efficient, and impactful solutions.

Our Tech Stack

The Agile RL logo
The AnyLogic logo
The Gurobi Optimization logo
The official logo for Red Hat OpenShift. It consists of a dynamic, red circular shape made of overlapping segments, creating a modern, glitch-like effect. Below the symbol, the word "OPENSHIFT" is written in grey, capital letters against a solid black background.
The PyTorch logo
The RLib logo
The TensorFlow logo
The Ultralytics YOLO logo

From Research to Reality

Our practice focuses on a portfolio of internal projects designed to build future capability. We explore, build, and validate, turning bespoke innovations into a robust, reusable toolkit.

Combining Autonomy (RL) with Language Models for enhanced autonomous decision-making, even on the edge.

Developing robust and adaptable autonomous behaviours. We build intelligent agents that make reasoned choices and act independently to achieve goals in complex environments.

Focusing on Causal Machine Learning and inherently Explainable AI to align with our robust AI TRiSM strategy.

Powering innovative and creative autonomous solutions. We leverage generative models to create novel training scenarios and strategies, accelerating the development of our agentic systems

We innovate with purpose. Our process is a commercially-focused pipeline designed to nurture ideas into high-value, reusable assets.

Ideas are sourced from across business and captured in a living backlog of research themes.

Our Steering Committee assesses potential ROI, strategic alignment, and feasibility.

Using a “fail fast” approach, we rapidly develop a Proof-of-Concept or Minimum Viable Product.

Successful prototypes are demonstrated and transferred as reusable “accelerators”.

Our Commitment

Responsible AI by Design

Our commitment to ethical AI is non-negotiable. We operationalise this through AI TRiSM (AI Trust, Risk, and Security Management), a robust governance framework that is a core pillar of our methodology and a key differentiator.

  • Explainability – Ensuring we can interpret and explain how models arrive at decisions, building crucial trust.
  • ModelOps – Rigorously managing models from deployment to retirement, with continuous monitoring for performance.
  • Security & Privacy – Implementing stringent security to protect models and data from threats, ensuring compliance and confidence.

Learn about AI TRiSM in our blog series.