Blog Post
Ushur XOS at the GenAI Summit
The week before last, Ushur attended the inaugural GenAI Summit at the Palace of the Fine Arts in San Francisco, CA. It was an incredible forum for speaking with and listening to the industry’s experts who are defining the future of Generative AI. At the summit, Ushur’s attendees also got the chance to ask questions about the latest release of the Ushur Customer Experience Automation (CXA) platform and its introduction of Ushur ExperienceOS (XOS) — an entirely new way to build customer experiences. XOS is the generative AI (GenAI) services backbone of the Ushur platform with a specific focus on using large language model capabilities to design engaging AI Agents for regulated industries, and to help Ushur users in the leading CXA platform.
XOS is the Generative AI (GenAI) service for launching AI Agents that help customers across several core competencies. It’s the backbone to GenAI services in the Ushur platform designed to bring the power of large language models (LLMs) to enterprise customer experiences. XOS hosts Ushur LLMs, open-source LLMs, or models trained and deployed by customers. Business users can use XOS to launch AI Agents — digital counterparts to the customer service representatives that customers love. Ushur AI Agents are trained to help customers and users in conversations over chat or voice, document processing, data report generation, and decision assistance for employees.
If Ushur users don’t want to rely on XOS to deploy AI Agents, they can leverage the power of LLMs via XOS in Ushur Studio. GenAI services help users as well as customers with tasks like workflow creation, phrase suggestions, and tone improvements to meet CX best practices. XOS is the operating system for all customer experiences that demand the fluidity of Generative AI, created with the ease of drag-and-drop. And, it’s available today!
One of the most common questions asked at the Ushur booth at the GenAI Summit in SF was “What are AI Agents?”
Artificial Intelligence Agents (AI Agents) by Ushur are digital customer service specialists that can be trained to adhere to a company’s industry or business use cases and automate processes for customer requests conversationally via chat, email, or voice, and securely gather and process documents for secure/PII customer experiences. These AI Agents are also able to help customer service agents make decisions via business rules and generate insights via data querying. Each of these capabilities are deployable out of the new Ushur GenAI service and accessible to customers via a conversational interface. AI Agents are the ultimate manifestation of GenAI in the Ushur CXA ecosystem.
To better understand AI Agents, it’s useful to think of them by comparison to their predecessor the chatbot. AI Agents powered by genAI can be trained and skilled to deliver insights, make decisions, and dynamically communicate at a higher order of reasoning and intelligence. Chatbots are generally pre-directed to respond with preset responses to specific FAQs that were accounted for. Reasoning and understanding are limited with traditional chatbots.
AI Agents have to complete end-to-end automation and drive towards straight through processing. They do of course also support agent escalation or diversion to a person if the customer needs (or prefers) it.
An AI Agent is as much defined by preconceived notions of related technologies as it is by anything else. They exist on a gradient of capabilities and are generally understood as the next iteration of intelligent assistants that have occupied the market for a decade. They’re more than rules-based chatbots in front of an FAQ set. Now, enterprises must decide what level of autonomy to give their AI Agents, and what skills their agents need in order to best deliver self-service to their customers.
Conversational capabilities are critically important for any agent that’s required to assist customers with self-service. Customers want to engage in the language or interface of their choice, and chatting has emerged as the de facto engagement style users prefer across all demographics. Chatting can be via text or web chat, but what’s clear is customers want to be able to speak to their service providers at their leisure and in their own verbiage.
Especially with the emergence of conversational capabilities of ChatGPT, the tolerance for poor chat interfaces has shrunk dramatically in the last two years. Fortunately, Ushur AI Agents skilled with chat use the power of LLM-based conversations to deliver a fluid and dynamic engagement but with the safety and compliance of Ushur’s industry-specific models and GenAI guardrails.
Customer experiences often depend on information already locked into a document-based format, and customers don’t want to repeatedly provide their data to companies ad nauseam. That’s why document processing skills for AI Agents are a requisite for any customer experience in the enterprise. Customers demand they be able to engage friction-free, and waiting for a customer service member to extract data from documents introduces immense friction.
With the advent of large language models, Ushur’s AI Agents are now better suited than ever to understand and correct the information included in the documents that customers share with their service providers. Forms (like pdfs, faxes, surveys etc.) are a mainstay of enterprise processes, and Ushur’s LLMs can be used to review and confirm the data extracted for those processes is accurate. Ushur can also use the power of LLMs to reach back out to customers when data in a document looks incorrect and ask for an update.
AI Agents are not just an improvement to existing SaaS platforms — like traditional machine learning projects have been – but are instead valuable in the ways adding another teammate to a customer service team would be. That’s why AI Agents that support self-service require skills like the ability to share data reports with other team members, to help other team members make decisions based on business practices, and to hold an initial dialogue with a customer over a channel like voice.
These skills are impactful because:
The news and internet are awash in examples of GenAI deployed without the kinds of bumpers that keep enterprise experiences on target. It’s a technology stack that is proving to be exceptionally compelling, while also a concern if bad actors can prompt-engineer those services into dangerous behaviors.
Ushur is specifically purpose-built to deliver secure, compliant, and closely guided experiences. Ushur services act as guardrails by validating all inbound prompts, and reviewing all generated answers to ensure they’re not off topic or serving up information via prompt attacks.
The unique value in the design of Ushur customer experiences is how they are goal oriented. That makes it easy for Ushur XOS and its GenAI capabilities to know when engagements are off topic, and then to guide users back through the experiences they’re intended to complete.
Ushur XOS has fundamentally changed the way that consumers engage with Ushur customers, and reimagined the process Ushur users have to go through to build customer experiences. The GenAI Summit was the perfect opportunity to validate those viewpoints, and show excited attendees that Ushur is thinking about how GenAI can make their customer experiences more powerful while preserving safety and security. If you want to see how AI Agents built via XOS are changing the way customers engage in the enterprise, come schedule time with us at ushur.com/xos/.
