Tier 1 + Tier 2 · Automation
Automation fails quietly. A trigger throws an error, the platform logs it, and nothing tells anyone; a workflow fails on every single run and the queue keeps accepting work; an integration is repointed at a test endpoint during a release and never pointed back. On one live tenant we found 33 of 152 active event triggers silently in error, one of them the trigger behind an automated conflict check. This audit reads the automation estate end to end and reports which of the things you believe are automated actually happen.
Every configured event trigger is checked for error state. A trigger in error raises no alert; the work it was meant to do stops happening, and the gap is invisible until someone asks why a step was missed.
Execution history per workflow: success, failure, and retry rates across the instance population. A workflow failing on 100% of runs across thousands of instances is not automation, it is a queue of undone work.
Integration endpoints resolved and verified, including the release-day classic: a production integration left pointing at a development endpoint, so live matters call a test system.
Instances stuck mid-flight: erroring, suspended, or untouched for months while the matter they belong to moved on without them.
95th-percentile latency per API route from platform telemetry, so the slowest routes your users feel every day are named and ranked rather than anecdotal.
Recurring error signatures grouped and ranked, and a specific check for personally-identifiable information (PII) being written into error logs, where it sits outside every access control you configured.
Trigger state, workflow execution history, and integration configuration are read through the administration and event-engine surfaces; latency and error telemetry come from platform diagnostics (Tier 2 scope). Everything is read-only. These checks are not yet part of the productised lens pipeline: they were built and proven on a live engagement and run today as engagement-scoped analyses, which is why every item above carries a roadmap chip.
On one live tenant, 33 of 152 active event triggers were silently in error, including the trigger behind an automated conflict check; several workflows showed a 100% failure rate across thousands of instances; and production integrations were found pointing at a development endpoint. The panel below presents those observed findings.
You learn which of the automations you rely on actually run, ranked by consequence, with a repair list your Sharedo team can action: re-arm the erroring triggers, fix or retire the always-failing workflows, and repoint the integrations before a missed conflict check or a passed date makes the gap visible the hard way.
