With 12,000+ registrations, the recent UiPath AI Summit virtual event served as a means for the AI automation company, one of the organizations leading the GenAI charge, to update users on the progress of technical developments at the company and reveal more of what’s to come.
UiPath delivered a series of announcements that collectively build a roadmap for how the company intends to accelerate delivery of GenAI technology to customers over the coming months. In this analysis, we’ll provide details on these announcements and how they should impact UiPath customers.
New Large Language Models (LLMs)
The most significant announcement from the summit was the launch of a new family of contextually specific LLMs that are unlike general-purpose alternatives. DocPATH and CommPATH are designed to tackle document processing and communications use cases, respectively.
DocPath will help businesses process any document out-of-the-box, including lengthy and unstructured data and tables. CommPath can process diverse communications, from transactional messages to complex requests containing multiple requests and context-specific language
Because DocPATH and CommPATH were trained for these particular tasks, they are more accurate and efficient in executing them than LLMs trained for a wide range of use cases. The new, customizable LLMs will be deployed by organizations to expedite document understanding and interpretation of communications, regardless of message type.
“Instead of relying on imprecise and time-consuming prompt engineering, UiPath DocPath and CommPath give businesses extensive tools to customize AI models to their exact requirements,” Graham Sheldon, Chief Product Officer at UiPath, said in a blog post discussing the announcements.
These LLMs are intended to give customers greater accuracy, faster time to value, and enterprise-scale controls with robust security, compliance, and governance, Sheldon added.
GenAI Activities
GenAI Activities are prebuilt functions that enable users to leverage a series of GenAI use cases including email text completion, data categorization, image detection, language translation, and personally identifiable information (PII) filtering.
GenAI Activities aims to dramatically reduce build times by providing users with out-of-the-box functionality that can be added to workflows. The technology should deliver benefits in areas including customizing customer experience, optimizing supply chains, and forecasting demand.
‘Context Grounding’ Makes Business Data LLM-Ready
UiPath introduced a new feature to its AI Trust Layer called Context Grounding that will help organizations improve GenAI models by using retrieval augmented generation (RAG) to extract information from business data.
Ask AI Ecosystem Copilot about this analysis
Ultimately, the process turns this data into an LLM-friendly format that’s easy to index, search, and use for more targeted prompts. Context Grounding will enhance users’ experience with UiPath GenAI across UiPath Autopilots, GenAI Activities, and intelligent document processing (IDP).
“Context grounding makes data LLM-ready by converting it to an optimized format (embeddings) that can easily be indexed, searched, and injected into prompts to improve generative AI predictions,” Sheldon wrote in his blog post. “AI responses will be more accurate, domain-specific, and significantly less prone to producing misleading outputs, sometimes called ‘hallucinations.’” Customers get more trustworthy models as a result.
Expanding UiPath Autopilot
UiPath has added to its Autopilot suite with Autopilot for Developers and Testers, currently in preview. This Autopilot taps GenAI to speed up automation, coding, and expressions during the development process. Autopilot for Testers is designed for testing teams to quickly generate test cases, generate automations based on test steps, and gather insights from tests in minutes without the heavy lifting that’s otherwise required.
The technology lets developers and testers create automations, apps, and Visual Basic expressions; convert legacy paper forms and PDFs to automation-powered apps; and get contextually relevant ‘next-best-action’ guidance when developing automations.
Closing Thoughts
In laying out this set of new products and initiatives, UiPath leaders noted the importance to customers of ensuring the business context of their AI models and applications, gaining access to diverse models that ensure high contextual value, as well as actionability with the AI models. These new products give customers a powerful set of tools to maintain that focus and deliver AI results in the near term.