Contract lifecycle management company Agiloft has launched a new product called AI Trainer, specifically designed to empower non-technical legal and business professionals. This innovative tool addresses one of the primary challenges that arise when using artificial intelligence software for contract review: training AI to align with a company’s unique preferences for phrases and essential terms. Agiloft aims to overcome this hurdle by enabling users to train specialized AI models that recognize key terms and clauses for their specific organization or industry.
No-code training capabilities in AI Trainer allow non-technical users, including subject matter experts and legal professionals, to train AI models on exclusive clauses and key terms specific to their organization. This bypasses the inadequacy of relying only on pre-trained, generic AI models that often fall short of meeting individual company needs. The new product provides legal, contracting teams, and other stakeholders within an organization, such as those in procurement, sales, and HR, with an easy-to-use, self-service tool to enhance their contract process’s automation.
A user initiates the process in AI Trainer by creating a training project for a particular clause to be extracted. The user then uploads an initial set of up to 20 documents containing variations on the preferred clause language, tagging and annotating each document to train the AI model. The system learns as the user proceeds, making suggestions based on preliminary inputs. Once trained on the initial set of documents, the model is published and becomes available to use in review workflows. Approximately a quarter of the documents are kept aside during the training process for testing the model and computing its precision and recall. The finalized model exhibits these precision and recall numbers.
The new AI Trainer has several potential applications. Among them, organizations can use it to review legacy contracts sitting outside the CLM platform and import those into Agiloft to apply the new AI model in identifying, isolating, and extracting vital clauses. Additionally, it can come in handy to review incoming agreement drafts from external parties, employing the AI models to pinpoint the key terms and clauses and benchmark them against standard templates or clauses in their clause library. An AI model can be trained to detect specific clauses within a company’s documents in response to changes in laws or regulations pertaining to their industry. Moreover, AI Trainer can also be used to search for specific clauses or language within contracts, such as force majeure clauses or specific data language.
Currently, Agiloft‘s model library can maintain multiple versions of models dealing with the same clause. This feature is highly beneficial when a company uses different language for indemnities in various agreements. The new AI Trainer is now available to Agiloft customers and requires an additional purchase for the AI module.
According to Andy Wishart, Agiloft’s chief product officer, the introduction of the AI Trainer aligns with the company’s broad vision of building connected, intelligent, autonomous, and self-governing contracting processes. It aims to help customers unlock the value of their contract data to accelerate decision-making and business growth. Andy Wishart adds that usability was a key focus in developing the AI Trainer, intending to cater to non-technical users within the Agiloft platform comfortably.
This release builds on Agiloft’s launch of ConvoAI in January, a user-friendly feature allowing users to interact with contracts in a chat-like interface. The AI Trainer brings the ability for organizations to utilize their best-of-breed AI in analyzing, reporting, and ensuring the effectiveness of contract management. This empowers close-knit teams involved directly in the contracting process, enabling them to train and customize the systems they use to uncover and categorize key terms and clauses in their contracts.
For more details, you can read more about it here.