Metric drift is the classic BI problem: two dashboards define revenue or win rate differently. Limnn solves it with a semantic layer of named metrics, and adds an advanced engine and Cortex Analyst on top.
Semantic metrics
A semantic metric is a named, reusable definition such as Pipeline Value, Win Rate or First Response Time: the object, the aggregation, the filters and the date field. Reports and dashboards reference the metric by name, so changing the definition once updates every tile that uses it.
- 1In the report builder at /reports/new, open the metric picker and choose an existing semantic metric where one fits.
- 2If none fits, define a new metric: choose object, aggregation, default filters and the date field used for time series.
- 3Name it in business language and add a short description so others reuse it rather than rebuild it.
- 4Use the metric across reports and dashboards and confirm totals against a known source.
The advanced engine
For questions beyond simple aggregation, use the advanced engine in the report builder: multi-object joins across the graph, calculated fields, period-over-period comparisons, cohorts and funnels. It still runs under the viewer's permissions and field access.
Cortex Analyst
- 1Open /reports/analyst.
- 2Type a question such as: which segment had the biggest drop in win rate last quarter?
- 3Press Execute. Cortex plans the query against semantic metrics, runs it and returns a chart, a table and a short narrative.
- 4Open the result's query details to verify which metrics and filters were used, then save it as a report if it is useful.
Insights
Cortex insight surfaces never run on their own. You type the question and choose Execute, so usage and credits are always deliberate. Results can be minimized, they auto-collapse after 15 seconds of inactivity and can be maximized for a closer look.