AI Training

The AI Training phase is a valuable option in the project setup process. Though not mandatory, it's particularly beneficial in certain contexts.

Application of AI Training:

Custom Terminology: Teach Athenic AI your specific language, such as unique KPIs or jargon like "target 1," to align its understanding with your company's terminology.

Clarifying Ambiguity: Explicitly define what subjective terms mean to your business, instructing Athenic AI on your preferred interpretation so it can respond accordingly.

How to Use AI Training:

Natural Language Training

Natural Language Training empowers you to instruct the AI using everyday language or the language of your choice. Guide the AI by providing specific commands or inquiries and associating them with database tables and columns.

Define Terms:
  • Formulate a command or a question as if you're speaking naturally. This could be a straightforward phrase or a detailed query.
  • Examples: Top Customers, Top Sales Season, What was my best sales day?

Detailing Your Definition:
  • Describe the task you want the AI to execute. Here, you will connect the earlier command or question to tangible data parameters like Table or Column Names from your dataset.
  • For instance, if you define "Customers" within a command: Clearly state that "Customers" refers to opportunities marked as "CLOSED WON.
  • Indicate the criteria for sorting, such as by the size of the opportunity.
Training Progress Bar
  • As you add more training key terms and definitions, you’ll notice the AI Training Coverage bar increasing
  • A fully trained table or column will be highlighted in green.
  • ** Remember, this training step is designed to fine-tune Athenic AI to your business needs, the more terms defined, the more accurate your queries will be.


SQL Training

To tailor your AI's understanding, map out your terms using SQL training.

Define Terms:
  • Enter the question or term you want to teach the AI in the left input field.
  • Examples: Top Customers, Top Sales Season, What was my best sales day?
Detailing Your Definition:
  • Using SQL define what action you want the AI to take each time it recognizes the specific key word.
  • For example, if you're training the AI to recognize "target 1s," your SQL query might identify these as opportunities with a probability greater than 0.7 and with "DO_NOT_CALL" set to false. Execute the query to ensure its accuracy, then click 'confirm' to store this knowledge pair. This process sharpens the AI's precision in understanding and responding to your specified terms.