Data ingestion & warehousing

To report on results alongside skill development, you first need the results in Admire. Artifact sources bring outside data in, from your CRM, support desk, or any system that produces records, and turn it into tidy, queryable tables that sit right next to your people and skills.

Each artifact source is a kind of record you track, with its row count and type at a glance.

Get data in

1. Create a source for a kind of record (deals, tickets, calls). It’s schema-on-read: Admire infers column types as data arrives and widens them automatically if new data needs it, so you don’t have to get the schema perfect up front.

2. Load the records, two ways:

   • CSV for a one-off or periodic load: upload a file and map its columns.
   • API for a live feed: push records in continuously as they happen in your systems, so the data stays current with no manual step.

   See Uploading & ingesting data.

3. Route the columns you care about. Each record is stored as arbitrary JSON, and a column routes a JSON path (even a nested one) to a typed value. Ingest rows of almost any shape and decide later which paths become columns.

Ingest with your AI

With your connected AI, you can describe the data you have and let it create the source and push records over MCP, which is handy for irregular or one-off feeds you don’t want to script.

Warehouse it: shape data with derived sources

Ingested data is rarely in the exact shape you want to report on. A derived source is a new source whose rows are produced by a SQL query over one or more existing sources, the place to build aggregations, cross-source joins, or computed tables. It materializes the result as its own queryable table, so you can chart it like any other source.

A derived source: the result of its query (here written in plain English) populates the new source. This is how you aggregate and join raw feeds into report-ready tables.

You can also add dynamic columns computed with SQL (a derived status, a bucket, a score) that store alongside your raw columns. Either way, the heavy shaping happens once, in Admire, without standing up a separate warehouse or BI stack.

With clean, queryable data in place, turn it into KPIs and analytics.