See the structure
Turn goals, information, and constraints into a clear representation.
Practise one problem-solving move, then try it where the surface has changed.
A Noesis workout teaches a named action—such as representing the problem, ranking constraints, or deciding when to change course. The next task changes enough to check whether you chose the action, rather than remembered an answer format.
The goal is not to finish more exercises. It is to find a better way into the next unfamiliar problem.
Training follows a progression—not an endless feed of puzzles.
Practise the move
Focus on one strategy until it becomes available without prompting.
Carry it forward
Use the same reasoning move when the story and surface change.
See one workout from beginning to end.
This example begins with the same meeting-schedule problem used on the Science page. The strategy is useful only if the learner can recognise when to use it again.
Starting pointThe learner begins assigning four meetings before listing the conditions, reverses course four times, and finishes after one prompt.
- Strategy card: list constraints before you calculate.
Name the move
Before placing anything, write down the conditions that remove the most options.
- Worked contrast: see why the second opening reduces rework.
Compare two openings
One solution starts filling time slots immediately. The other ranks the adjacency and ordering constraints first.
- Near practice: the structure stays; the surface changes.
Try a similar structure
The meetings become workshop sessions and the values change, but the same kinds of constraints still govern the answer.
- Independent check: support is recorded separately from success.
Remove the prompt
A second task asks the learner to choose an opening independently. No strategy name is shown.
- Transfer check: did the learner select the strategy in a new context?
Change the context
A delivery schedule now requires the same move, without the visual layout or language of the earlier tasks.
A recommendation should explain itself.
Noesis shows the observations behind every recommendation. For the attempt above, the explanation would read like this:
Why this strategy
Across three constraint-heavy tasks, calculation began before the conditions that ruled out options were identified.
Why now
The pattern appeared twice without hints and once after a revision. That is enough to prioritise practice, not enough to label a stable weakness.
What changes next
The next workout begins with one worked contrast, followed by two independent tasks with different surface details.
What would change our view
If constraints are identified independently in two different contexts, repetition is reduced. Contradictory evidence updates the recommendation.
Personalised training means three separate decisions.
Selection, challenge, and evidence checks are adjusted independently. Keeping them separate makes the next session easier to explain and easier to revise.
Selection
Which named strategy should receive attention now?
Choose “identify key constraints” after the pattern repeats; do not infer it from one wrong answer.Challenge
How complex should the task be, and how much support should remain?
Change the number of interacting conditions or remove a worked example—not merely make a timer faster.Evidence check
When should the system revisit the strategy on a different surface?
Assisted success schedules an independent task before the profile is updated.A curriculum for the decisions that happen before the answer.
Noesis organises training around reusable problem-solving moves—the kind that remain useful when the subject, numbers, or format changes.
Clarify the goal
Define what the problem is truly asking before choosing a method.
Filter the signal
Separate decisive information from detail that does not change the answer.
Build a representation
Turn words and conditions into a structure that is easier to reason with.
Decompose complexity
Break an unfamiliar problem into smaller decisions with clear dependencies.
Test the route
Check assumptions and intermediate conclusions before errors compound.
Switch with purpose
Recognise when an approach has stalled and choose a more productive one.
Your training history stays with the experience.
The core experience starts without an account. Strategy recommendations are generated within the product, and training activity is not sent to an external AI provider.
- Assessment and training activity stays on the device by default.
- Sign in with Apple and sync are optional.
- If sync is enabled, the app explains which records are uploaded and provides export and deletion controls.
- Removing local data without sync removes the history the on-device recommendation uses.
What progress looks like
- A named strategy appears more consistently on completed tasks.
- The strategy is selected independently after support is removed.
- The same move is recognised when the story, values, or layout changes.
- Later practice shows whether the new approach is retained over time.