Testing Strategy

← Back to index

---

I had limited time for testing. I therefore decided to familiarise myself with the capabilities of AI-Kernaro through simple and practically common use cases.

A. Working with AI-Kernaro directly in Enterprise Architect

1. Analysis of an unfamiliar repository
   1. Basic model element statistics
   2. Generating documents from a larger model
2. Creating a functional specification for a Coffee Machine
   1. Use case model with typical roles and scenario descriptions
   2. Activity diagram
   3. State diagram
   4. Generating a Word document with the UC model and its description
3. Diagram quality validation
   1. Identifying elements without a description

From a technical perspective, AI-Kernaro supports several collaboration modes:

1. Chat
2. Agents
   1. Internal
   2. Python
3. Documentation generation

B. Working with the Claude Platform Console

1. What is the approximate token consumption for different types of tasks?
2. What is the cost of individual actions – e.g. generating a Word document from a repository?
3. How can LLM costs be optimised?

---

Chapter	Topic
B – Chat	Natural language queries, document generation, TOC
C – Agents	Script Agent, Python execution, parallel agents
D – JavaScript	EA Script Manager integration
E – Findings	Consolidated test findings and recommendations

---

← Back to index | → B – Chat