The operator view is older than the dashboard, and better. A dashboard reports. An operator acts. Most platforms have spent a decade optimizing the wrong primitive — they pulled people away from the model and gave them a chart. Mulholland puts the model back in front of them.
The dashboard problem
Charts compress decisions into time-series. They flatten causes. They strip provenance. By the time you've read the dashboard, the data has already moved on, and your mental model is two layers removed from the row that triggered it.
The pattern repeats across every "modern" tool: a system of record buried under a system of presentation, with a layer of glue between them that breaks every six months and explains nothing on the way down.
A dashboard tells you what happened. An operator changes what happens next.
The ontology, fronted
Mulholland's bet is simple. If your business is a model — entities, relationships, derived state — then the operator should query that model directly, with the same expressive power as the engineer who shaped it. Not a curated subset. Not a "no-code" layer pretending to be one. The actual model, in Postgres, fronted by PostgREST, secured by RLS.
We pair that with Marzy: an agent that proposes the next move, asks for permission when it isn't sure, and acts when it is. Every action it takes is replayable, because it goes through the same model the human just queried. Auditors don't get a separate report. They get a different query.
What changes
Time-to-decision collapses. The "why" is one query away. The audit trail is the model's history, not a separate report. And the people closest to the work spend their day on the work, not on chasing dashboards that are stale by the time they load.
This is what we're building. It's why Mulholland exists.