Learning Intelligence
Create, deliver, assess and manage learning using AI-assisted workflows, grounded in learning science.
Work has changed
Most organisations can report on learning in detail — completions, certifications, training hours, attendance. What is harder to answer is whether all that activity is making people better at the work.
That question matters more than it used to. AI, automation and changing roles are reshaping work faster than traditional learning systems can respond, and the change is constant rather than occasional. Roles keep moving, and the next one may not exist yet. Hiring cannot keep pace, so capability has to be built rather than bought.
59%
of the workforce needs retraining by 2030World Economic Forum, 2025
Reporting has not caught up with it. It can show what was delivered and who completed it, but not whether the capabilities the strategy depends on are growing. Most leaders end up asking the same three questions.
What can our people do today?
What will the work need next?
Is learning actually building it?
85% / 7%
say building ability to adapt to change is critical — only 7% deliver itDeloitte, 2026
The gap is not effort, and not intent. Learning activity is easy to see; capability is not.
Build capability, not just learning
When work changes faster than organisations can keep up, learning has to start from what the business needs people to be able to do — and help each person build it as the work evolves.
Capability is the demonstrated ability to apply skills, knowledge and judgement to perform work. It is built one person at a time, and the organisational picture is what individual capability adds up to.
An example. An insurer moving claims to a new automated system needs handlers who can use the standard workflow and recognise when a case needs human judgement. Perpetua builds the learning from the insurer's own procedures and focuses practice on the objectives a learner has not yet demonstrated.
Across two thousand handlers, those gains are how the organisation becomes capable of running the new model.
What Perpetua LIS does
Perpetua LIS creates the learning, personalises it, and measures the capability it builds.
Version 1.0 delivers the first. It is pilot-ready now, with general release in Q4 2026. The other two are what we are building. Everything sits on one data model, so what you use today is a complete product in its own right and every stage that follows builds on the same foundation.
Adaptive Intelligence
Learning that adapts as each person progresses — practice, scenarios and role play, shaped by how they are doing.
Capability Intelligence
Workforce capability modelled from what the business needs, and measured from the evidence learning produces.
What is in 1.0
Everything below is in 1.0, and is demonstrated from the real product.
Course design and curation
Build a course from your own source material. The system asks the questions a learning designer would.
Assessment, built in
Generated alongside the content, and every question traces to a learning objective.
Quick Checks
Standalone assessments you create and run on their own, outside a course.
Narration, video and publishing
Produce narration and video, then publish, without leaving the system.
PF Learning Assistant
Answers each learner in the context of their own course material, as they work.
Readiness Check
Sets each learner's starting level so the course meets them where they are.
Analytics you can question
Ask your own learning data questions in plain English, rather than building a report.
Enterprise administration
Roles, permissions and controls for running learning across an organisation.
See it working
A short preview — how a course, its assessment and its evidence are produced from a document you already have.
How it is built
It is built from your own material. A course comes from a document your team already uses, or from a prompt — not a generic library.
The assessment is built with the course, not after it. Authoring and assessment run together, so every question traces to the objective it tests and nothing has to be collected from elsewhere.
Intelligence runs through the whole system. Specialised agentic AI works in sequence across design, content, assessment and analytics, rather than sitting on top of one part.
Capability starts from the business, not from job titles. What people need to be able to do comes from what the business is trying to do — not from a role description or a CV.
Evidence here means the output of assessment, practice or observation. Never completion, attendance or time spent.
How the AI is governed
Perpetua drafts. People review and publish.
Nothing reaches a learner until a person has reviewed and published it. Learner identity is never exposed to the AI, and an organisation's material stays in its own environment. Security, privacy and responsible AI are designed in, and we are working toward ISO/IEC 27001, ISO/IEC 42001 and SOC 2.
Where this leads
Strategy→Capabilities→Skills & gaps→Personalised learning→Measurable capability→Business outcomes
Each step defines the one after it. What the business needs determines what people must be able to do; that determines what they learn; and the evidence they produce is what capability is measured from.
It ends at business outcomes rather than at learning, because that last step is what separates capability from training. We call this capability-led learning — learning designed from what the business needs.