Every leadership team already has AI ambition, and capable models are no longer the constraint. What's missing is the ability to turn a plan into a system that actually runs inside your existing teams. We are a high-impact engineering-led consultancy addressing this need.
Projects need someone who can see where AI will actually create value, and someone who can build and operate it.
When either is missing, projects stall or solve the wrong problem. We bring both: the experience to grasp your business's complexity, and the technical know-how to build new capability without tearing down what already works. You own everything we build.
The opportunity we advocate for has nothing to do with adding AI assistants. Your team already feels the confidence they gain by using LLMs. The real value comes from connecting the tools, data, signals and decisions across teams, so the whole system of how your business operates jumps to the next level.
A chatbot in one window and a CRM assistant in another. Each is useful on its own, but none of them are aware of the others. People type a little faster while the business stays as fragmented as before.
Data, workflows and decisions are joined up across departments, on top of the systems you already run. Sales calls inform support, and support informs product. The system improves as it is used, and the business moves faster than competitors who are still stitching tools together.
We're clear on delivering outcomes that have unmistakable revenue impact, with a practical edge.
We start with commercial judgement, finding where AI will create measurable value and setting an ambitious but realistic plan.
We build production systems rather than demos: LLM applications, statistical and vision models, and the architecture that supports them.
We put the system into your infrastructure and in front of real users, keeping it auditable, governed and measured against the outcome it was built for.
We provide documentation, code and knowledge transfer, so your team can own, run and extend what we built.
We work to tight timeframes. The aim is useful systems in the hands of their users within days and weeks, so that AI is adopted at the pace the field is moving. If your internal projects are taking multiple quarters, we offer a faster alternative. Most engagements start small too, one workflow or one team, not a company-wide rebuild. The scope naturally expands when data sources reveal their potential for further AI transformation.
A mix of Cambridge AI Works engagements and the founder's own track record building and running production AI.
Before founding Cambridge AI Works, Michael spent two years as the sole AI lead inside a SaaS revenue organisation, building and running a suite of LLM-powered tools now used daily across sales and customer success: deal-qualification scoring, pipeline forecasting, win/loss and churn analysis, prospect research, follow-up drafting, customer health monitoring, and an upsell recommendation engine. Every feature shipped from idea to daily use within days or weeks, one capability at a time, backed by the same discipline we bring to client work now: backtested models, audited scoring, and a human in the loop before anything reaches a customer.
Michael architected and built visual search using CLIP-style embedding models and vector databases, so images can be found by their content without relying on metadata. The same work covers task-specific and object classification, on-premises video scene detection, and semantic similarity. It runs on GPUs in a horizontally scalable AWS cluster, using worker queues, distributed processing and autoscaling that hold up under production load.
Alongside commercial AI, we write scientific, engineering and data-science software, and build and operate specialist data platforms at real scale, including a database platform used by a global user base of more than 100,000 people. The common thread is engineering discipline: statistical correctness, reproducibility, and systems that stay reliable long after launch.
A specialist toolkit spanning strategy, applied research and production engineering.
We would rather build something that works than present what might be possible. Here is how that plays out on an engagement.
Cambridge AI Works is a highly specialised consultancy. You work directly with the people who build your system, rather than through a layer of account managers. Engagements start at whatever scale makes sense, a single workflow or a full platform, and you deal with us directly from day one.
Michael has 25 years' experience building and running SaaS businesses, including a successful merger of Third Light Ltd to PhotoShelter Inc. in 2022, where he subsequently bootstraped an AI engineering team and AI revops transformation project.
He works across commercial and technical foundations, combining AI strategy, deep learning, LLM applications for revenue and operations, and the cloud and GPU systems that run them in production.
LinkedIn
Nick focuses on AI, advanced engineering and technology innovation. He founded and led Global Inkjet Systems as CEO and CTO, building the company into an international industrial-technology business before its acquisition by Nasdaq-listed Nano Dimension, where he subsequently served as CTO.
A named inventor on several industrial-technology patents, Nick specialises in building and commercialising complex technology products. His experience spans business automation, product development, engineering software and the application of AI to scientific and industrial problems.
LinkedIn
Tell us where the pressure is in your operations. We build and operate a focused system to relieve it, and most start delivering within weeks. A typical first engagement is scoped to a single workflow, runs two to four weeks, and hands you a working system as well as next steps.
Cambridge AI Works