1. Evaluate
System, configuration, knowledge and results.
Cognitive diagnosis
Discover what is limiting your AI's performance.
We evaluate your existing system, identify recurring failures and review its knowledge, professional role and possible causes to propose a clear, actionable plan.
◎ Evidence-based analysis◇ Focus on your context▥ Practical priorities
FUNCTIONAL ANALYSIS
No client dataCognitive Diagnosis
From symptoms to a testable hypothesis
Does it reliably handle equivalent tasks?
To be evaluatedDoes it retrieve relevant, up-to-date sources?
To be evaluatedWhere does it need review and why?
To be evaluatedDoes it follow procedures and limits?
To be evaluatedSelect an area to discover how we investigate it. Methodological guidance, without data from your system.
We reviewEvidence
We testHypotheses
We deliverPriorities
2. Analyse
Recurring failures, patterns and possible causes.
3. Prioritise
Impact, risk and effort in relation to your objective.
4. Propose
An improvement plan with a defined scope.
What we investigate
Beyond the symptoms.
Accuracy and consistency
Functional quality, variation across runs and recurring errors.
Knowledge and sources
Relevance, currency, contradictions, retrieval and information traceability.
Role, procedures and tools
Role definitions, resource use, restrictions and operational limits.
Human oversight
Which corrections people make and where they need to intervene.
Evaluation and risks
Representative tasks, success criteria, critical errors and regressions.
Cost per useful task
Time, resource consumption and review work required to obtain a valid result.
Preliminary self-assessment
What is limiting your AI?
Do not enter documents or confidential data. This guidance does not replace a professional diagnosis.
From uncertainty to a clear plan