Analytics
See what your agents do, and what they cost
Every conversation leaves a trail. We turn it into dashboards you can actually read — volume, resolution, latency, and LLM spend, broken down per agent, per channel, and per day — then let you ask Saba what changed in plain language.
What the analytics suite shows you
A dashboard that reads at a glance
Conversation volume, resolution rate, active agents, and total cost sit on one overview — with the trend against the previous period so you see direction, not just a number frozen in time.
Drill down per agent
Open any agent to see its own conversation count, message volume, resolution rate, and spend. Compare your Support bot against Sales and spot the one that's quietly underperforming.
Drill down per channel
Telegram, WhatsApp, Slack, the web widget, and the rest each get their own breakdown — so you know where your traffic actually comes from and where the cost is concentrated.
Trends over time
Every metric plots as a time series. Watch volume climb after a launch, catch a latency regression the day it starts, and see whether a prompt change moved resolution up or down.
Cost and LLM spend, tracked to the request
Every model call is priced and rolled up, so usage and credit burn are never a surprise. See spend per agent, per channel, and per day — and know exactly what's driving the bill.
Ask Saba what moved and why
Skip the chart-staring. Ask in plain language — "which agents are winning this week?" — and Saba reads your own analytics and explains the shift, the same answer it gives inside the panel.
From conversation to insight
Analytics is automatic — every conversation is measured and priced as it happens, then aggregated so the dashboards, drill-downs, and Saba's answers all read from the same numbers.
Every conversation is measured
As agents talk, message counts, resolution outcomes, and latency are captured automatically — no instrumentation to set up, no events to wire.
Every model call is priced
Each LLM request is costed against current model pricing and attributed to the right agent and channel, so spend rolls up alongside usage.
Metrics are aggregated
Raw activity is rolled up by agent, by channel, and by day, and scoped to your organization — so the numbers stay yours and stay fast to query.
Dashboards and drill-downs render
The overview, per-agent, per-channel, and time-series views all read from the same aggregates, so what you see lines up no matter where you look.
Saba explains it in words
Ask a question and Saba queries the same analytics to summarize what changed and why — turning the dashboard into a plain-language answer.