ROI Analytics
ANTS Platform measures the cost side of AI return on investment and the savings you can get from better model choices. It does not hold revenue data or calculate an ROI percentage for you; combine its cost and savings figures with your own business value numbers.
The views below need Gauge AI read access (Owner, Department Admin, Security & Compliance and Cost Analyst). See Members & Roles.
Cost per unit of work
| Question | Where to look |
|---|---|
| What does one run of an agent cost? | Gauge AI → Cost Dashboard → P 95 Cost per Trace and Top 20 Use Cases (Trace) by Cost. See Cost & Latency Dashboards. |
| What does each user cost? | Top 20 Users by Cost, and Gauge AI → Users. |
| What does an agent cost per day and month, and at scale? | Discover AI → AI Agent Center, agent detail, Cost tab. See AI Agent Center. |
| Which model gives the best quality per dollar? | Gauge AI → Cost Performance Metrics, model costs next to scores. See Cost & Performance Metrics. |
Savings from model choices
The AI Optimizer tests cheaper models on an agent's real traces and reports:
- Cost savings per trace and as a percentage for each comparison model that keeps quality.
- Projected monthly and annual savings on the Recommendations tab, based on a fixed volume of 100,000 requests per day.
- On the Analytics tab, total projected savings, a monthly savings trend, optimization status and per-agent analysis for the last 30 days.
See Recommendations for exactly how savings are calculated.
Putting business outcomes next to cost
To relate cost to value, record the outcome of each interaction as a score, for example a BOOLEAN score named resolved or a NUMERIC customer rating. Scores appear in Cost Performance Metrics and Latency & Analytics next to cost and latency, and can be grouped in a custom dashboard.
Next steps
- AI Optimizer: produce savings you can measure.
- Cost Tracking: attribute cost to users, teams and environments.