Alias:

Aliases are a big part of training. Aliases map business terminology to database columns. For example, mapping "revenue" to orders.total_amount. This is how Teela learns your company's vocabulary. Admins can create and edit aliases.

‍

Client:

This is what Teela calls a company. Each Client can have many users and many connections.

Connection:

A connection is literally a connection to a database. Teela supports MySQL, SQL Server, and Postgres databases. Depending on the size of your business, you may have one to dozens. For each connection you will need to know the host, port, database name, user name, and password. Don’t know how to connect? No problem, just invite your IT coworker to help you out and we will take it from there.

‍

DataClip:

So you have your query ready to go. Now what? Now you save it as a DataClip (also known as a DC). You can save it just for you, or you can share it with your team. You can even schedule it to email you every Thursday. DataClips are the backbone of Teela. The cool part is that once saved, DataClips are regenerated whenever you need and they are super fast.

It is also important to know that a DC is a foundation that you can build upon. Lets say you have a great query saved as a DataClip. A week later you run it, but you want to refine it even more. No problem! You can just run the DC, then ask additional questions about it to change it up even more. If you like what you have created, save another DataClip!

Also, if a DC is timely and you want to see it regularly, you can set it up to automatically run and email you the results on the schedule you choose.

Dashboard:

Lots of apps have dashboards, but Teela lets you build one that is completely unique to you. All you have to do is create the perfect DataClip and then call it up to be part of your Dashboard. It can be a table, a single number, or a beautiful chart. You can make it small or huge. You can even set it up to alert you if the results change too much. Customizability is the name of the game.

‍

DDL:

On the training side, you will see DDL. This means Data Definition Language, but practically speaking this is how Teela keeps track of your database structure. Once we have saved all of your database information (not data), we save your tables as DDLs. What is important is that from this moment on, Teela does not go back to your database to know its structure. We use these DDLs to tell us what your tables are and how they are related. In other words, if a table or column isn’t in the DDL, then Orko has no idea that it exists. This can be a good thing if you want to hide tables and columns for security reasons.

DOC:

Doc is just short for Documentation. In Teela training these are your best friends because they are where you can tell Teela how your business works in your own words. No technical jargon needed. For instance, if your business calculates revenue in their own unique way, this is how you tell Teela about it. Most training that you will make will be a DOC.

‍

Folder:

These are simple ways to keep track of your DataClips. Folders can be private to you only, or you can share them with everyone on your team.

Orko:

This is what we affectionately call our custom query engine that powers Teela. Orko is the beating heart of the system. All natural language questions are processed and turned into intelligent (and often quite complex) SQL queries that allow us to consistently retrieve data using AI without hallucinations.

‍

Query:

Technically this is just a question, but in Teela it is all about getting the exact question that you want. “Show me the top 10 customers based on revenue in the last 3 months”. That is a great question, but that isn’t the end. 

Teela returns you a nice table, but maybe it has too many columns, or is showing you data you don’t care about? Now it is time to refine it. “Remove XYZ column” or “change it to 6 months” or “show me customers only in the Northeast”. The true magic isn’t in the first question. It is working through the query in your own words until you get the exact results you want. That is what we mean by a query in Teela.

Principle Query:

The Principle Query (often shortened to PQ) is the first question that is asked in a new Thread. This becomes the focus of the entire thread. For instance, if you ask “Show me all of our customers in Idaho”, then this becomes your PQ. From here we understand the context that you are asking about user information so you don’t need to mention that with your follow up questions. With that said, if you suddenly start asking about invoices, then Teela will get confused and may ask you to start another Thread.

‍

Roles:

All users are assigned one of three roles: Viewer, Client Admin, or Client Owner.

• Viewers can run queries, save DataClips, and create a Dashboard. They basically can use Teela to answer questions about their data.
• Client Admins are able to train Teela and adjust what is and is not visible to Orko when asking questions.
• Client Owners can do everything a Client Admin can do, but they also are in charge of billing.

Schema:

This is the overall structure of your database. Think of a schema like a blueprint for how your data is organized inside the database. It includes the tables, columns, and relationships that make your data unique.

‍

SQL Query:

In the database world, this is a unique question written in the SQL language, but in Teela it is also a training type. If you understand SQL you can save it as training to let Teela know exactly how to answer a question. Another way to save SQL data is when you have a great question that shows exactly what you want, Admins can save it for training. You simply explain what the SQL is answering and then future questions can benefit from this training.

Thread:

‍A thread is a conversation that a user has to get the right results. You will see threads under the History header. Unlike other AI chatbots, the goal of a thread is not to come back to it over and over. The goal is to get a query refined such that a DataClip can be saved to use in the future. In fact, when a DataClip is saved, the thread is archived because its job is done. You can get it again if you need to by clicking the dots next to History and viewing the archives.