Press Releases

Barista LLM Gateway for Safe ChatGPT Use

Share this post:

The Barista Employee Language Cloud is the First Domain-specific LLM for Employee Language to Safely Leverage the Massive Content of a Generalized LLM

SANTA CLARA, Calif., March 21, 2023Espressive, the pioneer in automating digital workplace assistance, today announced the new Espressive Barista Large Language Model (LLM) Gateway™, enabling its virtual support agent (VSA), Espressive Barista, to offer enterprises safe and responsible use of large language models (LLM) like ChatGPT. Espressive Barista serves as a gateway to LLMs that organizations can utilize to enforce policies protecting privacy, such as not permitting the sharing of source code or personal identifiable information (PII), as well as providing the ability to “turn off” LLM access for areas such as Epic Systems in healthcare organizations or HR content. The newly expanded Espressive Experience Selector™ will prioritize automation and customer content versus a generative response to ensure employees follow best practices recommended by their organizations for compliance. Additionally, new advancements in the Barista Control Center combine conversational AI, employee context, automation, and prompt engineering to unlock the full power of ChatGPT and other LLMs while delivering even faster time to value with an improved employee experience.

“Generative pre-trained transformer (GPT) is a truly transformative technology,” said Gaurav Garg, founding partner at Wing Venture Capital. “Combined with virtual agent technology, it has the potential to significantly save costs, accelerate productivity, and help open a world of opportunity—and empower more people to do their best work.”

Unlocking a New World of Workplace Assistance and Productivity While Minimizing Privacy and Security Risks

Espressive started building its highly secure domain-specific large language model (LLM), the Barista Employee Language Cloud™, over 6 years ago. Specifically focused on the employee self-help use case, today it is trained on over 4 billion phrases across 15 enterprise departments in over 100 languages. It incorporates technologies that enable industry-leading accuracy and deflection rates with a layer that is dynamically tuned to understand the nuances of an organization’s lexicon. The Employee Language Cloud includes automations and integrations, and combines customer knowledge as potential outcomes versus only providing a generative AI response.

By adding integration to generalized LLMs in a safe and secure way, autonomous resolution rates will increase, further driving down the costs of employee service management. In addition, new workplace and employee productivity capabilities will be made available since Espressive Barista will be able to answer an even wider range of questions from how to run highly effective hybrid meetings to understanding what the conversion rate should be on the organization’s website. Further, with Barista integrations to third-party systems such as Zoom, employees could request a list of their recent Zoom meetings, meeting summaries, and action items.

“Although CIOs see ChatGPT and other LLMs as extremely promising, they are struggling with the potential loss of PII or corporate secrets from employees sharing this information in their interactions,” said Pat Calhoun, founder and CEO of Espressive. “Enterprises, such as JPMorgan Chase, have gone as far as denying access to ChatGPT services from corporate resources. With systemic controls and support for company-specific policies, customers can unlock the full power of LLM services in a secure manner, without having to worry about the disclosure of sensitive information.”

The New Espressive Barista LLM Gateway

Integration with generalized LLMs like ChatGPT provides Espressive customers with the advantage of a domain-specific LLM for automating the resolution of employee questions combined with safe access to the massive knowledge of a generalized LLM. Espressive advancements include:

  • Barista LLM Gateway: Instead of simply allowing integration to LLMs, Barista enables customers to enforce a set of policies to control how LLMs are being used. For example, an employee can ask about the best restaurants in the area, but not be able to ask how to best diagnose a patient. Additional safeguards include the ability to verify policy compliance (e.g., no source code, no PII, no data with customer names); disable access to LLMs for specific content areas (e.g., Epic or HR); restrict questions to only those that are work related; and ensure employees receive notice that responses from LLMs may not be completely accurate.
  • Experience Selector: With the Barista LLM Gateway, the Experience Selector can choose a response from an LLM such as ChatGPT. It uniquely considers relevance and possible outcomes to provide the best experience to employees and prioritizes automating resolution and customer content versus a generative response. When a less than optimal outcome for employees is identified, it uses Smart Ticketing to create, attribute, triage, and route a ticket to the right service team, minimizing mean time to resolution (MTTR).
  • Barista Control Center: With LLMs, the response is only as good as the question, and getting the best outcome requires employees to be experts in prompt engineering, the process of designing prompts or inputs to an AI language model to elicit specific responses. New advancements in Barista Control Center eliminate this complexity by combining conversational AI, employee context, automation, and prompt engineering to provide customers with the confidence that Espressive Barista will leverage known information and integrations to deliver the best possible outcome to their employees.
  • Translation Engine: Espressive was the first to support translations in over 100 languages, and with the Barista LLM Gateway, customers can now also leverage LLMs for translation services based on the best possible translation engine to use for a given topic in a given language. For instance, Barista can use Lionbridge for Latin based languages, but ChatGPT for Japanese and Baidu for Chinese.

For more information on Espressive Barista request a demo here.

About Espressive

Espressive, the pioneer in automating digital workplace assistance, and a 2022 Forrester New Wave Leader, redefines how employees get help by delivering exceptional employee experiences. Barista, the company’s virtual support agent (VSA), brings the ease of consumer virtual assistants, such as Amazon Alexa and Google Home, into the workplace. Barista automates resolution of employee questions, issues, and requests with personalized experiences that result in employee adoption of 80 to 85% and reduced help desk call volume of 50% to 70%. Espressive is backed by Insight Partners, General Catalyst, and Wing Venture Capital, and is headquartered in Santa Clara, California. Visit www.espressive.com for more information.

Media Contact

Pam Njissang
Bhava Communications for Espressive
espressive@bhavacom.com