Find the operator workflows worth automating.
Prove which ones aren't.
EXOS turns real OpenTelemetry traces into operational process mining, product analytics, and session replay — all on ClickHouse, all SQL-native. It surfaces the high-frequency, high-determinism workflows that are ready to automate, and the bespoke ones that genuinely aren't.
Automatability is earned from response determinism given the extracted inbound state — not asserted from a bare received → sent sequence. Illustrative readout.
See the same OpenTelemetry data three ways
The browser RUM and backend spans land once in ClickHouse. Each lens asks a different question of them — from "what should we automate?" down to "show me the exact session behind that anomaly."
Cluster on intent × extracted-entity completeness × response determinism. Each cluster earns an automatability score from how concentrated its outbound response is — with honest non-automatable counterexamples, not blanket "automate" labels.
Explore process miningOrder-lifecycle funnels, conversion and drop-off, operator engagement and retention cohorts — computed over the real RUM and trace data, cross-linked so you can pivot from a funnel step straight to the sessions behind it.
Open analyticsHundreds of thousands of real browser RUM spans, enriched with operator identity. Jump from a multivariate z-score anomaly straight to the recording — the exact clicks, scrolls, and dead-ends that produced it.
Watch sessionsReal traces in. SQL-native analysis out.
No black-box pipeline and no fabricated datasets — the data is synthetic, but the path from raw span to insight is the real one you'd run in production.
Browser RUM plus backend spans, modeled to OTel semantics — resource attributes (telemetry.sdk.*, service.version, deployment.environment), span kinds, realistic latencies, and clustered error skew.
Every span lands in a columnar store on GCP. There's no pre-aggregated cache standing in for the data — each panel hits otel_traces directly. Hover any metric on the dashboard to see the exact query behind it.
Process mining, lifecycle funnels, and anomaly detection (a multivariate z-score at 2.5σ) all run as SQL — auditable, reproducible, and portable. The analysis is the query, not a vendor's secret.
Synthetic data, real analysis
The traces in this demo are generated, not captured from production — so the numbers will differ from a live deployment, and nothing here is a customer's data.
What is not faked is the analysis: there are no Math.random() fallbacks and no fabricated result sets — every figure is a live ClickHouse query you can read.
The synthetic corpus is modeled to be structurally indistinguishable from production OpenTelemetry (spec-faithful resource attributes, span kinds, latency models, and error distributions) so the mining and analytics features have something legible and realistic to operate on.