Challenge · Glass Box

We turn hard-to-read AI decision logs into a review screen people can actually follow.

This Glass Box view pulls in the challenge logs and explains, in plain language, what the system did, why it stepped in, and which cases may need a human to look closer.

Last event · ...
Observed events
...
All of the log entries we loaded into this review screen.
Blocked actions
...
Cases where the system stopped an action instead of letting it go through.
Escalations
...
Cases the system pushed toward more review or human follow-up.
High severity
...
Cases marked as more serious and worth extra attention.

What this page helps you see

What people can understand here

What happened

Each case shows what the system did and when it happened, so nobody has to read raw logs first.

Why it happened

We surface the rule, reason, and suggested correction that led to the system response.

Why this matters

This helps judges, reviewers, and teammates quickly spot risky cases and decide whether a human should step in.

Important note

This Glass Box demo is currently powered by simulated audit data.

We are using challenge-supplied sample logs to demonstrate how this review layer would work in a real production system. In Holmes, a surface like this could help teams catch AI hallucinations, spot when data refresh jobs fail or run with stale inputs, trace why a model response was blocked or corrected, and flag cases that need a human to step in before decisions reach people in the real world.

That matters because this platform touches sensitive civic information that could inform housing decisions, connectivity planning, and public-facing explanations. If a system like Holmes is going to be trusted by industry teams, city partners, or operations reviewers, it needs a clear audit trail that explains what happened, why it happened, and when someone should investigate further.

Event Playback

Recent inhibitor timeline

Click any recent event to have Holmes explain what happened in simple terms.

Loading event…
...

Loading the latest challenge log traces.

Loading event…
...

Loading the latest challenge log traces.

Loading event…
...

Loading the latest challenge log traces.

Loading event…
...

Loading the latest challenge log traces.

Loading event…
...

Loading the latest challenge log traces.

Audit Breakdown

Actions and policies that surfaced

Click an action or trigger below and Holmes will explain what it usually means.

Action share

Most common triggers

Holmes Explainer

Click a timeline event to get a simple explanation

Selected event

Select a recent timeline event to inspect it.

Holmes summary

Click a recent timeline event and Holmes will explain what happened, why the system stepped in, and what a reviewer should take away.

Holmes Breakdown

Click an action or trigger to learn what it means

Selected summary item

Select an action or policy trigger above to inspect it.

Holmes explanation

Click any action bar or policy trigger card above and Holmes will explain what it means in simple terms.

Expected input

Read-only CSV and JSONL inhibitor logs. We do not modify the provided sample_logs baselines; we ingest and render them.

Rendering behavior

The dashboard parses event timestamps, actions, severities, and policy triggers into summaries, timelines, and reviewer-friendly inspection views.

Why this helps audit review

A compliance or demo reviewer can quickly answer what happened, why the intervention occurred, and what should be checked next without reading raw logs.