Intelligence
The Intelligence portal: its GPT, Reports, and Data sections, and the six GPT tabs.
The Intelligence portal gives your team situational awareness of the activity occurring on your equipment, and shows how your copilots are performing in the real world. It combines data gathered by InSkill with information from integrated third-party systems — maintenance history, current operating state, and asset history — which InSkill also uses within its large language model (LLM) operations to optimize the experience for you. It turns raw interaction data into actionable insights, letting you identify knowledge gaps, track engagement, and continuously refine your system's expertise.
The Intelligence portal's sections and GPT tabs
This system data not only informs the LLM but can provide critical insight into questions you or your team might have. Analysis of the data gathered by InSkill can help paint a better picture of what is happening in your system — whether it is slow wear and tear of the machine or an efficiency miscalculation — and can help diagnose an issue that arose in your testing.
The portal is organized into three sections, with tabs in the top-left corner of the page: GPT, Reports, and Data.
GPT
The GPT section is the core of the Intelligence portal. It tracks your team's use of InSkill GPT, the chat interface where Technicians get real-time help, so you can monitor the effectiveness of its conversations. InSkill GPT mines its information from the expertise loaded into the copilot, and needs an internet connection. The section is broken into six tabs:
| Tab | What it shows |
|---|---|
| Summary | A high-level dashboard of key metrics: volume (chats and questions, including a questions-over-time chart), engagement (active users over time, chats by role or language), quality (thumbs-up responses and feedback received), and technical detail (questions by copilot, references by type, and agent usage). |
| Details | A granular breakdown of every individual session, including date, user, question, answer, references used, and feedback provided. |
| Questions | Clusters similar questions together to reveal common inquiries across copilots and roles. |
| Answers | Clusters similar answers to help authors verify consistency and quality across responses. |
| Issues | A repository for negative interactions — thumbs-down responses, flagged content, and No answer and No references cases — for admin review. |
| References | Lists every GPT agent, task, knowledge base, and resource InSkill GPT used, with how often each was used, helping identify the most and least valuable documentation. |
Reports
The Reports section gives you an easy-to-understand summary of what is happening on your machine. Information on sessions, resources, copilots, users, and customers is gathered and presented through tables, graphs, and charts, and can be segmented by user group, copilot, and time period.
Data
The Data section provides a more in-depth look at your machine. While reports are shown as easy-to-understand visual aids, data is presented as a spreadsheet with detailed records on users, customers, copilots, sessions, customer sessions, resources, customer resources, and account-level activity. You can narrow it to a specific time frame, choose which columns to display, and export it.
Updated about 16 hours ago