Avvoka's Smart Automation feature transforms static Word documents into dynamic templates by automating placeholders, conditions, and Docusign tags based on specific syntax being used in Word, allowing users to "automate the automation".
Smart Automation v2 enhances this functionality by using AI to more effectively identify and apply automation to the document.
How to access SmartAuto v2
When creating a template, this new "SmartAutomation Ai" tab in the creation modal will be available for selection, containing the various smart automation settings and options.
How it works
Placeholder Detection
If your placeholder text is contained within square brackets in the word document, smart automation will convert it into an Avvoka placeholder. It is also directly capable of applying transformations to the placeholders based on context such as UPPER CASE, lowercase, Capitalised, and Sentence case.
All placeholders detected are afterwards postprocessed by the LLM to generate an appropriate attribute name and question. Each question also has a type that best represents the real value of the placeholder (e.g.: numerical values = number type question). The LLM can also merge multiple placeholders together if they are found to refer to the same value.
Condition Detection
Conditions are also detected with square brackets as well with LLM Classification. The system scans the document for all remaining bracket pairs and evaluates whether the text is truly conditional or not. In addition to the text itself, any footnote following the text is passed on to the LLM as additional context and guidance. All conditions that were detected in the document are postprocessed by LLM to generate an appropriate name and question. Please note that currently all conditions are constructed as a 'select' type question.
Additional Options
Users are able to upload commercial data to help guide the LLM for proper automation. Term sheets and emails help aid LLM decision-making when automating.
Ignoring any prior saved AI responses allows LLM to refresh and take new automation in to account. Although token consumption is increased, this yields better results in automation.
E-signature tag detection also allows Smart Automation v2 to detect tags that may be used by Docusign or other e-signature platforms.
