Analytics
Steer your company’s AI: usage, costs, satisfaction
A dashboard to decide, not just observe: who uses AI, on what, with which models, at what cost, and with what satisfaction. You set the caps up front, you do not discover the bill afterward.
Control
Spending set in advance, never discovered after the fact
Set a cap per person, per agent and per model, by day, week or month. In strict mode, calls stop at the cap. In soft mode, you are alerted at the threshold and at the cap, and the service continues. Inference stays with your provider, with your keys, no margin.
- Cap per user, per agent and per model
- Period of your choice: day, week, month
- Strict or soft mode, alert at threshold
- Multi-provider BYOK, no markup on tokens
Usage
Who uses what, with which model
Track consumption per user, per agent and per model, over the period of your choice. Spot the most used models, the busiest channels and the most engaged teams, and adjust your choices with full knowledge.
- AI usage per user and per agent
- Most used models
- Breakdown by channel: app, Teams, Slack, widget
- Active users and conversation volume
Usage per agent, this month
- Product support46 %
- Sales and quotes31 %
- Legal and contracts23 %
Quality
What your teams look for, and what they are missing
Questions cluster by theme and keyword: you see what your teams look for. Answers without documentary support are flagged, so you complete your sources where it matters. Satisfaction surveys close the loop.
- Questions clustered by theme
- Most frequent keywords
- Answers without documentary support flagged
- Satisfaction collected at the end of a conversation
Question themes
- Warranty and returns
- Delivery times
- Product compatibility
- Billing
- Access and passwords
3 questions without documentary support this week
Analytics questions
- Are analytics included in the Enterprise AI Platform offer?
- Yes. Usage, spend, questions, satisfaction and channels are part of the platform. The sealed audit log belongs to the governed offer.
- Can we cap without cutting service?
- Yes. In soft mode, an alert fires at the threshold and at the cap and service continues. In strict mode, calls stop at the cap.
- Do analytics data leave the platform?
- No. They live in your instance, SaaS or self-hosted, and are viewed in the interface.
Want to see GuardWeaver at work in your company?
A thirty-minute demo with our team, on your own use cases: your rules, your tools, your teams.