Admit it: You’ve always wanted to have a conversation with your contracts. Well, now you can, thanks to ConvoAI, a feature introduced today by Agiloft, the contract lifecycle management company, during its Agiloft Summit 23 conference for users and partners in Las Vegas.
The feature lets users interact with contracts in a chat-like interface, asking questions in natural language to more easily find details in their contracts. It is designed to be used by anyone within an organization, not just legal or contract professionals, Agiloft says, and can be accessed either from within Agiloft’s CLM platform or its Microsoft Teams app.
Andy Wishart, Agiloft’s chief product officer, said that ConvoAI provides a new way for organizations to unlock information in unstructured contracts.
“Our new ConvoAI feature not only enables contract operations professionals to search and find information more easily, but it will also help these professionals drive greater adoption of CLM across the enterprise by providing business users with an intuitive way to query the contracts they are responsible for, and find the data they need, to speed value creation in their organizations,” Wishart said.
ConvoAI will come standard for all new and existing customers and is accessible from the Agiloft CLM user interface and Agiloft for Microsoft Teams.
Agiloft says that two features, in particular, make ConvoAI unique for searching contracts:
- It learns more quickly. ConvoAI uses Cognizer’s cortex neural vector engine that includes large-language AI transformer models initially trained by reading huge knowledge bases and millions of commercial contracts, and then fine-tuned on hundreds of contract types, clauses, and terms across 500,000 annotated samples. Compared to CLM AI products that use prior generation neural networks, ConvoAI is already “smarter,” and can learn new clauses and terms with fewer examples and less human effort. As new types of clauses and key terms are added to the models, the types of questions ConvoAI can answer will grow.
- It thinks differently. Genius stores and organizes data that it learns in a knowledge graph, database technology that creates connections and relationships among related data and concepts, which means the model can answer natural language questions about all of the contracts in the knowledge graph, such as how many contracts have a particular clause or attribute. This is different than other CLM AI solutions, which extract and populate data in fields in a standard database, preventing these solutions from answering natural language questions, because the data is not mapped in a way the system can retrieve the required information.
“Combo AI enables our users and our customers to have a conversation with their contracts in natural language,” Wishart told me in a briefing earlier this week. “We’re making it easy for all users of Agiloft, not just contract operations professionals, but all users, the people working in the business, to be able to get to the thing that they want in the way in which they’re thinking about.”
Wishart said users will be able to directly ask questions such as, “How many contracts contain a termination for breach clause?”, “What is the termination for breach provision in a specific contract?”, or, “How many contracts have a force majeure clause?”
The feature will have a chat-like interface in that, after a user asks a question, a search pane opens in which the user can ask follow-up questions to refine the results.
As noted, the feature is also available in Agiloft’s Teams app, which Wishart says will make it easily accessible for managers and others in the business outside of contract professionals.
Although the application is driven by APIs, customers’ contracts never leave the Agiloft managed services platform, Wishart said.
“That’s really important for our customers and it gives us an approach that really simplifies things from a data privacy perspective,” he said. “Customers’ content never leaves that Agiloft hosted managed services.”
ConvoAI will be available beginning in March for its U.S. customers on its cloud managed services, and then shortly afterward it will be rolled out to customers in the U.K. and Europe.
Even though this new feature is driven by advanced AI, Wishart said that Agiloft’s overarching goal is not to focus on AI, but rather on what it can do for its customers, such that the AI is effectively hidden.
“Our vision for AI within Agiloft is that AI will be everywhere as we execute on our long-term vision towards fully autonomous and self-governing contracts — AI will underpin a lot of this. But what is really critical to us is not to make the AI the thing. The AI is just supportive of what the customer is trying to achieve. So AI should be everywhere but at the same time it should be nowhere, in my opinion.”