Connect Your Databases & Build an AI Semantic Model in Minutes

0:00Step 1: Connect a SQL Server operations database

First, we add a new data source. We name it Casino Operations, and choose SQL Server. Next, the host, the port, and the database name. Then the login, always encrypted. Connected! Seventeen tables and a hundred fifty-eight columns, mapped automatically.

0:25Step 2: Connect a PostgreSQL finance database

Now, the Casino Finance database, for the financial dashboard. It runs on Postgres, so we just enter the host and the database name. Add the credentials, and connect. Done! Sixteen tables and a hundred twenty-six columns, ready to analyze.

0:45Step 3: Build the semantic model with Vexon

Connected tables are just raw data. A semantic model teaches Semavex what that data means for your business. We open the Casino Operations model, and switch to edit mode. We could pick tables by hand... but why do the work? Instead, we ask Vexon, our AI modeling agent. We describe the business in plain English: who we are, what we report on, and how we define our numbers, like win, hold and occupancy.

Send it, and Vexon gets to work. It reviews every table and samples real values. That takes a few minutes... so let's skip ahead! And here's the proposal! Vexon wrote the business context for us: our three casinos, the currency, the gaming day, and every definition. Every table and column gets a plain-language description, with a confidence level, so you know exactly what to double-check.

Sensitive data, like loyalty card numbers, is flagged automatically. Vexon also works out how the tables connect, and lists anything it wants us to confirm. It even explains its assumptions, and adds performance tips. No black box here. One click to apply... and boom! Seventeen tables, a hundred fifty-eight columns. The business context now lives in the model, and every column has a role: a metric to add up, a dimension to slice by, or a date for trends.

And the measures? Vexon built them straight from our business goal. Table Win, Table Hold percentage, formula already written. Columns can also carry synonyms, so GGR and gaming revenue land on the same number. Twenty-eight relationships connect visits, slots, tables and venues, so one question can span the whole casino.

And here's the whole picture, as a diagram. Save, publish... and the model is live! From now on, every question, report and alert runs through this model. So everyone gets the same numbers, in the language of your business.

2:47Step 4: A second model for finance

Now, a second model, for Casino Finance. Each data source gets its own semantic model, speaking its own business language. Same recipe: we give Vexon the finance goal. Revenue, costs, profit, cash flow, and supplier payables. Vexon gets to work... and once again, we skip the wait. Same proposal as before: business context, descriptions, and its assumptions, all open for review.

Apply... Ten tables, a hundred and three columns. Vexon even left out the tables this report doesn't need. Column roles and finance measures are already in place, built from the goal. Fourteen relationships connect journals, accounts, and payments. Save, publish... and Casino Finance is live, right next to Casino Operations.

One last thing: Vexon accepts your input in any language, not just English. Describe your business in Spanish, French or Portuguese, and it works just the same.