Enrolment
This course is currently free to enrol in.
Public course
Enterprise AI — Course 4: Improving AI Task Performance
This course develops practical decisions about AI improvements, targeted experiments, model choices and when to retain, narrow or stop a workflow.
AI improvementcontextinstructionsmodel selectionadaptationtoolsexperimentsscope
By Telari Labs
Learning on Telari
Created by Telari Labs, delivered through Telari
The course creator controls the material and assessment standards. Telari provides structured lessons, practice, feedback against those standards and a record of your progress. See how Telari works.
AI access
The first managed lesson is included on Free. Continue with Plus or your own supported API key; premium assessments and certificates may require Plus.
Audio
No reviewed lesson audio is currently advertised for this course.
Lesson list
- Diagnosing before changing the systemA useful improvement starts with evidence about the failure and a change that can test an explanation.
- Improving instructions and examplesInstruction and example changes should repair a defined behaviour without narrowing the task or reinforcing an unsupported conclusion.
- Repairing information accessInformation repair should address the source actually needed and distinguish access problems from poor use of available evidence.
- Decomposition and deterministic toolsSeparating interpretation, deterministic work and checks can help, provided the interfaces and added work are evaluated.
- Model choice and adaptationModel and adaptation choices should follow task evidence, available data and the responsibilities of maintaining the chosen approach.
- Running improvement experimentsA bounded experiment sequence should resolve useful uncertainties and reveal which changes deserve to be retained.
- Stopping and narrowing scopeFurther optimisation should be justified by its incremental benefit and a workable scope, with clear reasons to reopen the decision.
- Choosing and challenging an improvement planThe final assessment combines intervention choices, experimental evidence and decisions about continued use.