Agiloft has announced the latest release of its contract lifecycle management (CLM) platform. Release 26 includes new generative AI capabilities that enable organisations to further automate the processes within managing contracts. Enterprise Times spoke to Andy Wishart about the release.
The Generative AI Prompt Lab
Wishart spoke about the main thing in release 26 is the generative AI Prompt Lab. This new feature enables subject matter experts to build templated prompts that can then be reused across the organisation. Ultimately, it will free up time for users to perform more valuable work.
Wishart explained the new feature, saying, “We’re really excited about Gen AI prompt lab. We think this is the first of its kind within the CLM space. It enables our customers to take any data within our data first platform, to infuse that into a prompt template, take the output and store it within Agiloft and use it within any workflow.
“With the launch of Gen AI Prompt Lab, we’re going to be shipping with five prebuilt templates that cover use cases.”
- The creation of a simple description of the contract
- A more detailed summary of a contract
- The creation of a simple termination letter for a contract
- A description of the comparison between two clauses and between versions of a document
- A summary of the supporting documents for a contract
Wishart continued, “Each of these pre-built prompt templates is editable or they can create new prompt templates as well. Over the next few releases, we’ll be continuing to release more prebuilt prompt library templates as well.”
Once built, the Termination letter contract could be reused across any contract. It would automatically extract the key data points from any contract. Ensuring that the letter is personalised to each specific contract.
Jack Wicks, Senior Product Manager at Agiloft, “GenAI Prompt Lab embodies Agiloft’s distinctive, flexible, data-first philosophy. By offering the versatility to leverage our generative AI models or to integrate your data with your own models, we ensure that, as AI technology advances, your workflows evolve in tandem. GenAI Prompt Lab truly delivers personalized AI your way.”
Gen AI Prompt Lab works beyond Agiloft
The press release infers that customers can also bring their own LLM models. I asked Wishart to explain about this feature.
“I can confirm that the Gen AI prompt lab does enable our customers to either use the connectivity that we provide or they can connect to their own accounts running in Microsoft Azure, or OpenAI to connect to their ChatGPT 3.5 or 4 models. We’ve worked with customers on this concept and this has been really important to them. Because they’ve been experimenting within their organisation with using Open AI models for contract and use cases. They may have created some fine-tuned models that are unique to their organisation, or they may have some custom terms in their agreement with Microsoft or Open AI.
Wishart then explained how customers can take advantage of this, saying, “They can take a copy of our pre-built templates and connect to their models. They can either adjust the template or perhaps their models are providing slightly different output compared to hitting the generic models that we are connecting to.”
Wishart concluded, “We’re enabling the builders within our customers and our partners to extend use cases across the platform where generative AI can be used. That’s why we think this is really special. We believe we’re the only CLM that can infuse the data in a very flexible way, with large language models, taking the output and using that to power the next stages of the workflow.”
Other features
The release contains two other major updates for customers. The first are enhancements to the integration with the Google ecosystem, notably Google Docs. In release 25, Agiloft added a contract assistant for Google Docs. In this release, the functionality for Google Docs is further enhanced. Bringing some of the features available for Microsoft applications to Google Docs. Google users can now access the Agiloft clause library from within the application. It enables them to analyze, markup, and reach agreement on third-party documents all from within Google Docs.
Agiloft has also enhanced its AI search capability. ConvoAI Document Q&A enables users to ask questions in natural language about their contract repository. The new feature enables business users to extract useful information from their contracts with simple questions. Eliminating the need to seek answers from legal teams.
Wishart explained the tool, saying, “Convo AI, is a new experience within the Doc viewer of Agiloft that enables the user to ask questions about an individual document and that integrates with a GPT model. It’s an extension to the Comvo AI natural language search concept that we announced back in February of last year. We’ve got further things in the future around this concept of natural language search.”
I asked Wishart whether users could change the tone of the language in the responses to suit a specific role. He replied, “It is something that we’re looking at. But for this release, it’s just a single experience, irrespective of your role within the organisation.”
The final major improvement is an improvement to the data synchronisation between the Agiloft contract record and the Word document associated with it. If the Gen AI Prompt template or a business change initiates a change within the Agiloft contract record, then those changes are reflected within the Word Document. The synchronisation is bi-directional so that if a Word document is changed during negotiation, the Agiloft record is updated.
What is next
I also asked Wishart what Agiloft is working on next. He replied, “Over the next few releases a continuation of new templates within the Gen AI Pompt Lab. We’re going to see further extensions to our AI-powered search. Using AI to power the findability of contracts and how users will interact with the CLM using natural language. We’ve got some other use cases that our team are researching and applying, prototyping to solve some really complex problems with large hierarchies of contracts that have been amended over time. So more to come.”
Wishart also sees an opportunity for customers and partners to create and make templates available outside their organisations. While Agiloft has not yet built a marketplace, it is exploring how to let the greater community share their templates.
Enterprise Times: What does this mean
This is another substantial release from Agiloft with improvements that take the use of Generative AI to the next stage. The improvements to the integration with Google Docs also make the platform more attractive to a wider audience.
What will be interesting is seeing how partners, especially, take to the Gen AI Prompt Lab and build their template libraries. They will either be able to use these to offer better service to specific verticals. Or, if they are able to productise them to a wider audience, improve the strength of the Agiloft ecosystem. If enough partners build libraries, Agiloft may need to accelerate plans for a marketplace.
It is perhaps disappointing that Agiloft users are not able to change tone or language in responses. This is often seen in other vendor solutions. However, building responses that offer business language rather than legal phrases should be something they look to do sooner rather than later. There are, however, risks with simplifying responses, and Agiloft will have to be cautious about how it applies language tone filters for specific legal terms.