Built on over 15 years at the forefront of AI and learning

Built by people who had to make learning provably work.

The platform has been in market since 2018, built on more than a decade of our founder's enterprise learning and consulting experience, and it grew up inside one of the hardest environments there is to deploy anything: banking.

Proof first, platform second.

The methodology came before the software. Years of enterprise consulting kept surfacing the same problem: organisations spent heavily on development and could not show what it changed. The answer became a discipline, understand the role, measure the person, aim the learning, prove the improvement, and from 2018 that discipline became a platform, first deployed to serve a major bank's learning solution and then generalised. Academic institutions have been among our clients and collaborators along the way, and the framework itself was built and validated in-house, against real performance data. Today the platform runs in production in financial services, with 50,000+ people assessed and developed on it.

A practitioner who was building enterprise AI before it was a headline.

Dr Juan Swartz

Executive Director

PhD in Strategy (with Organizational Behavior and Finance) Chartered Accountant (CAANZ) CGMA (UK)

Juan has spent more than fifteen years building and deploying enterprise AI inside large organisations, across banking, aviation, telecoms and insurance. He was putting AI-enabled operating models and predictive systems into production years before ChatGPT turned the technology into a headline. A Chartered Accountant with a PhD in Strategy, and a former Chief Strategy Officer of South African Airways, he reads the commercial case and the operating model as fluently as the technology itself.

That mix is rare. In a market now flooded with AI experts who arrived in the last five years, Juan brings the one thing that cannot be improvised: real enterprise deployments that shipped, through every shift in how the technology has evolved. He knows both the business and the arc of AI intimately, not just this moment of it.

What drives him now is people. Juan built AI Unlox on a conviction that the organisations who thrive in an AI-shaped world will be the ones that prepare their people deliberately, with learning aimed precisely at the capabilities each role and each person actually needs, and proven against real performance. Targeted learning, not generic training, is how he believes businesses and the people inside them stay ready for what is coming.

The enterprise AI pedigree behind the platform.

The team that builds and runs the AI Unlox Platform has spent more than a decade building and integrating AI and enterprise systems inside large organisations, in banking and financial services, property, telecoms, research and the public sector. That work is where the platform's engineering standards come from, and it is why the AI inside the platform is not a bolt-on.

A production multi-agent architecture

The platform's AI runs on the team's own multi-agent architecture, engineered for production: agents that analyse data, assemble auditable reports, orchestrate workflows across systems, execute tasks and coordinate stakeholders. That is the machinery behind the coaching, the learning assistant, content creation and the interpretation of results.

Model agnostic, by design

The platform uses best-of-breed AI models through an abstraction layer, with no lock-in to any single provider. When a better model arrives, the platform gets stronger without a rebuild.

Governance engineered in

Access control, PII tokenisation, audit trails and encryption are architected into the platform, not added later, with dedicated tenant isolation per client and in-country hosting for data sovereignty. That is what lets it live inside a bank.

Cost optimisation, built in

Because the platform is model agnostic, each task runs on the most cost-effective model that does it well, and the whole platform is engineered to optimise token usage. AI running cost is treated as a design constraint, not an afterthought, and token usage is passed through at cost.

Microsoft Azure Kubernetes Multi-region resilience Model agnostic Token usage optimised Governance by design Dedicated tenant per client In production in financial services

All of it serves one product: the AI Unlox Platform and the individualised learning journey it delivers.

Plain words, honest claims.

We say what the algorithm does

The individualised journey is algorithm-driven. AI does the coaching, the assistant and content creation. We will not dress one up as the other.

We publish what we can defend

Our stats come with their sources, our validation is our own work against real performance data, and client names appear only with the client's sign-off.

We build on what you own

Your LMS, your content, your data. We add the intelligence layer, and your identifiable data stays yours.

Talk to the people who built it.

A discovery call is with Juan, not a sales team. Bring your hardest question about measuring capability; that is the conversation we enjoy most.

Book a discovery call

How it works The science How we work