The AI hype cycle seems to have left many customers in the ditch of disillusionment. Everlaw may have avoided that roadmap trap with a heavy investment in a year-long beta of their EverlawAI Assistant. 2,900 beta users in 125 companies got free, unlimited use in exchange for regular feedback and debriefing sessions. Everlaw continues to invest in proving their value proposition with a free trial of the AI Assistant now that it is GA. I covered the Everlaw AI Assistant in my Legal Week briefing with AJ Shankar and recently got an update from him covering the beta cycle before the release.
eDJ Questions and Everlaw responses:
What AI enhanced features made GA?
- Coding Suggestions – help users find relevant or responsive documents faster with less effort. Based on user-provided instructions, AI helps prioritize and classify whether codes should be applied to documents. Coding Suggestions enables workflows that can meaningfully reduce manual human review, while maintaining review quality. In Everlaw’s tests with a real world customer dataset, coding suggestions achieved a precision score of 0.77 and a recall score of 0.82. On recall – a measure of how many responsive documents are successfully found – this beat first level human reviewer metrics by 36%. In addition, Coding Suggestions achieved high accuracy, correctly suggesting codes 88% of the time and when not to apply them 83% of the time. This enabled the team to confidently prioritize and filter based on these recommendations.
- Review Assistant provides a deeper and more comprehensive understanding of docs faster than manual review alone. With the click of a button, AI instantly generates document summaries and topic, sentiment, and entity extractions across a document set.
- Writing Assistant leverages AI to quickly understand key evidence and generate fact analysis in memo, outline or list formats. It can be used to brainstorm, like a thought partner, pulling directly from the key evidence in a matter helping craft an argument or build a counter argument while citing key documents inline. Finally, it can be used to refine those arguments or other content written in Storybuilder.
How do customers consume AI Assistant services without sticker shock?
- Clear Pricing – The customer controls usage at the database level and overall for the organization. They will also see the cost before they execute any generative AI task. Usage is not prescriptive to the customer. Want Review Assistant summary of that long Teams chat or contract? Here is the cost before you hit go.
LLMs are adapting quickly, how does Everlaw keep up?
- Everlaw has kept up with evolving models and AI pricing. AI Assistant has flexible API connectivity to leverage the best model for the most efficient/effective functionality. They are actively developing total corpus models like RAG with a focus on reliability, but most of the beta updates focused on prompt engineering based on user feedback for maximum value. Testing showed that smaller, private models performed poorly for recall functionality.
How does Everlaw protect customer data used for summarization or analysis?
- Everlaw only uses models under the Eurodata retention policy. Customer data is never used to train models or retained beyond the function call. All AI connections are role based and time bound (i.e. just-in-time access).
What kinds of usage cases were seen in the beta cycle?
- The beta customers comprised a wide range of industry roles, organization size and experience levels. The early adopters focused on downstream review and case analysis usage cases. Everlaw has reduced ECA pricing and expects utilization to work upstream as customers better understand capabilities.
Even at Legal Week I thought that Everlaw’s bounded approach to leveraging LLMs felt more practical and usable than the ‘Better, Faster, Easier’ AI banners I saw plastered over too many booths. As I recently commented on LinkedIn, “Everyone was excited about the potential, but the products have not yet been tuned to solve pressing pain points with a clear value prop. Just plugging GPT on top of existing workflows is not going to revolutionize customers’ lives. Some smart cookie is going to apply this tech to solve a real problem with a completely fresh approach.” While I am not convinced that the EverlawAI Assistant is a ‘completely fresh approach’ it does address real eDiscovery review problems with a clear value proposition and transparent cost.
Greg Buckles wants your feedback, questions or project inquiries at Greg@eDJGroupInc.com. Book a free 15 minute ‘Good Karma’ call if he has availability. He solves problems and creates eDiscovery solutions for enterprise and law firm clients.
Greg’s blog perspectives are personal opinions and should not be interpreted as a professional judgment or advice. Greg is no longer an investigative journalist and all perspectives are based on best public information. Blog content is neither approved nor reviewed by any providers prior to being published. Do you want to share your own perspective? Greg is looking for practical, professional informative perspectives free of marketing fluff, hidden agendas or personal/product bias. Outside blogs will clearly indicate the author, company and any relevant affiliations.
Greg’s latest nature, art and diving photographs on Instagram.
[instagram-feed num=1 imageres=thumb showfollow=true