Editorial illustration
Why do we need MOBILAR® if we already use ChatGPT or Claude?
During several MOBILAR® demos, a recurring question came up: why use a specialised platform when a general AI assistant is already available? It is a fair question. A general model can explain high-voltage battery diagnostics and suggest development topics. To answer which technicians in your network need support, it also needs access to the right profiles, assessments, workshop context and evidence. The interface alone does not supply that foundation.
Connected does not automatically mean meaningful
Retrieval-Augmented Generation (RAG) helps AI retrieve relevant information before answering. The Model Context Protocol (MCP) can connect an AI interface to specialised systems and tools. These approaches can make company information accessible, but a connection does not define how technicians, tasks, skills and assessments relate. Those relationships need to be modelled and maintained.
The intelligence behind the answer
A useful answer starts with the work a role is expected to perform, the skills that work requires and the evidence available for each requirement. Assessment results and observable behaviour need to remain distinct from repair-support signals, course participation and unknowns. Relevant information must also be current, accessible only to the appropriate people and traceable to its source. These are questions of architecture and business expertise, not just prompting.
Why retrieval quality matters
Sending thousands of unrelated records to an AI model can increase token usage, latency and cost without improving the answer. Good engineering narrows the question, retrieves relevant evidence and controls what is processed and stored. Reliable access and predictable costs matter when a conversational experience serves many users. There is no single model or prompt that removes those design responsibilities.
One interface, a specialised foundation
Imagine using ChatGPT as a conversational interface connected to MOBILAR® as a specialised platform. The model provides general reasoning and dialogue; MOBILAR® provides structured relationships between people, tasks, skills, assessments and development needs. This illustrates how the two can complement each other, not a claim that every named interface is already integrated. The MOBILAR® Agent makes this foundation easier to explore.
Keep evidence and judgement visible
MOBILAR® has spent years building the underlying models and workflows. AI helps make that intelligence more accessible; it does not replace it. The Agent prepares proposals, experts validate requirements, assessors confirm practical evidence and people decide on action. The important question is not whether AI can answer, but whether the knowledge behind its answer makes it relevant, checkable and useful.
A practical example
The Agent proposes that a technician may need support on camera calibration, based on support cases and a self-assessment. It labels this as inferred and suggests a practical assessment; the assessor then confirms or rejects it. (Illustrative.)
Questions we often hear
Does the Agent decide who needs training?
No. It prepares proposals with their evidence. People decide.
What does the AI use as reference?
Your role profiles, tasks, required competence and the evidence you connect.
