Built with your experts, not instead of them
Your specialists stay in their own job. Their knowledge gets structured, sequenced and instrumented by someone whose job that is, through a process designed around the time they can realistically spare.
Learning design for subjects where the detail has to be right.
Ambai Learning builds learning systems for organisations whose subject matter is technical, regulated or specialist. Universities and awarding bodies, training providers, and international development organisations.
Not courseware on its own. The pipeline that produces it, the competency framework it answers to, and the evidence that comes back out of it.
Scroll↓
What you get
The pipeline that makes the content, not only the content
A course is a deliverable. A pipeline is an asset. What gets handed over is the system that produced the material, the rules it runs on, and the documentation to keep using it without us.
Instrumented at design time, so completion stops being the measure
Competency mapping and xAPI statements specified while the learning is being designed rather than retrofitted afterwards, so the data answers questions somebody actually asked.
Most consultancies hand over a specification and leave you to find someone to build it. This designs the system and builds it. The large agencies win on scale and studio production; Ambai Learning works where the subject is difficult, the audience knows the material better than any writer, and the output has to survive scrutiny.
The working method
The model never gets the final say.
AI is used heavily here, and it is used under supervision. The distinction is not decorative: it decides whether what you receive is defensible when somebody senior asks why a decision was made.
- The system does
- Drafting at volume, first-pass structuring, variant generation, candidate competency mappings, localisation drafts, consistency checks across a large body of material.
- The specialist does
- Writes the design rules the system has to satisfy, decides the sequence, verifies accuracy with your experts, and signs off. Every output traces back to a stated principle rather than a prompt.
Not a human in the loop, which implies somebody checking at the end. A human who set the constraints before the loop began.
Where this work fits
- Universities and awarding bodies
- International development organisations and multilaterals
- Professional and technical training providers
- Standards bodies and competency framework owners
- Publishers moving into structured digital learning
- Anyone holding a large library they cannot prove works
The common factor is not the subject. It is that the subject belongs to your people, the audience is expert, and procurement expects standards compliance rather than a showreel.
What the work is
Three service lines, done properly, rather than everything done thinly.
Curriculum architecture for subjects that resist being taught as topics
Whole-task sequencing, outcome sets that survive validation, and assessment that measures the thing rather than the recall of it. Built with your subject experts, who keep their day job rather than becoming instructional designers.
Systems that report what was learned, not what was opened
xAPI and cmi5 pipelines, LRS architecture, LTI integration, and competency mapping in CASS JSON-LD against a published framework or a bespoke KSA set. Specified, built and handed over with the documentation to run it without us.
Content pipelines with the design rules written into them
Instructional design principles encoded as machine-readable constraints the system has to satisfy, with a specialist reviewing what comes out. Your team gets the pipeline and the rules, not a dependency on an agency to run them.
Swipe or scroll →
The method
Instructional design encoded as machine-readable skills, kept under a human who can be held to it.
Most AI content tooling puts the model in charge and hopes for the best. This does the opposite: design principles are written down as explicit constraints the system has to satisfy, and a specialist reviews what comes out. Every design decision gets taken at three levels, and the constraints are written for all three.
How engagements work
Work is scoped as a defined piece with a defined output, not an open retainer. Most engagements start from one of three positions: a programme that needs restructuring, a content pipeline that needs building, or an existing library that needs mapping to a competency framework.
Everything is specified against open standards and handed over with the documentation to run it in house. The measure of a good engagement here is that it ends.
- StandardsxAPI · cmi5 · SCORM · LTI 1.3 · CASS JSON-LD
- Design models4C-ID · SOLO · Merrill · Mayer · Cognitive Load Theory
- Competency mappingAny published framework, or a bespoke knowledge, skills and abilities set
- DeliveryRemote, based in Madrid, working across European, African and Asian time zones
- AccessibilityWCAG 2.2 AA audit and remediation of existing libraries
- HandoverSource files, documentation and the pipeline itself, licensed to you
Start a conversation
Tell us what you are trying to teach, and who is struggling to learn it.
Scoping conversations are free and usually short. If the work is not a fit, we will say so and point you toward someone it does fit.