Tickets Contained
Man Hours Saved per Month
Reduction in Call Wait Time
.avif)
.svg)
Everise, a leading customer experience company, partners with top brands in industries like healthcare, logistics, and travel. With over 20,000 employees worldwide, Everise partnered with Retell AI to integrate voice AI within their own operations, transforming the efficiency of its internal call center.
Using Retell AI’s voice AI solutions, Everise transformed the power of its own call center
Tickets contained?: 65%
Man hours saved per month: 600
Reduction in call wait time*: 100%
*call wait time reduced from 5-6 minutes to zero
“We were able to contain 65% of voice calls with the bot, which previously would have been directly gone to a manual agent for resolution.” --Saurabh Sodhani, SVP of Digital Transformation, Everise
Everise specializes in delivering seamless, scalable, high-quality customer experiences for its clients, with 90% of its business touching voice channels. With over 20,000 employees, Everise saw an opportunity to test voice AI within their own internal global service desk before proposing and deploying it as a client solution. Integrating Retell AI boosted the efficiency of their service desk and saved invaluable man hours.
As SVP of Digital Transformation, Saurabh Sodhani recognized the transformative potential of voice AI for Everise’s customer experience voice channels. To thoroughly evaluate its impact, Sodhani decided to pilot the solution internally within Everise's global IT help desk. Everise partnered with Mindcraft Labs to integrate Retell’s voice AI solution. Utkarsh Marwaha, Co-Founder of Mindcraft Labs, outlined the project requirements:
Mindcraft Labs set out to design an intelligent “middle layer” between Everise employees and service desk resources. The voice AI solution would have to seamlessly integrate with Everise’s internal systems to efficiently identify callers and direct their call. Employees could be calling from anywhere in the world, including externally.
Everise’s previous solution relied on a complex IVR system, which required employees to navigate multiple branches to access service desk support for their particular issue. To replace the IVR, the voice AI would have to be able to recognize a caller’s identity as well as their intent.
Any solution needed to be able to identify whether the user had already created a ticket for their issue and create a new ticket when needed. The system would ideally be able to solve caller problems, redirecting the call only when absolutely necessary.
Everise employees called in from various countries, including the US, India, Guatemala, and the Philippines. The solution needed to cater to the different language preferences and accents of this diverse global team.
Through intensive UAT sessions with various global stakeholders, Mindcraft Labs set out to overhaul Everise’s IVR and replace it with an intelligent voice AI solution that could solve caller problems and minimize the need for human support. Retell AI provided the ideal solutions.
Mindcraft Labs were able to quickly code APIs to integrate Retell AI’s voice bots with Everise’s internal systems. Since Everise uses Workday to manage employee information, Mindcraft Labs used Retell AI to pass a caller’s employee ID into Workday to qualify and identify callers.
“Once I gave my Workday ID, I would describe my case and the voice bot, thanks to its NLP abilities, understands my intent and then branches me into the right place.” Utkarsh Marwaha, Co-Founder of Mindcraft Labs
Retell AI’s voice bots understand caller intent using Natural Language Processing (NLP). By integrating with ServiceNow, which Everise uses to manage internal IT tickets, Retell AI’s voice agents could recognize whether callers had already created a ticket in ServiceNow or whether a new ticket needed to be connected.
“The voice bot, thanks to its NLP abilities, understands caller intent and branches the call into the right place.” Utkarsh Marwaha, Co-Founder of Mindcraft Labs
Mindcraft identified six use cases, including account access, software issues, and telephony issues. The team designed a series of multi-tree prompts in Retell AI to identify how the caller’s intent mapped onto these issues, with each action connecting to its own API. The bot solved 65% of caller issues without needing to transfer the call to a human agent for further assistance.
“It's not just one single bot. It's like a squad of bots, each handling a different department coordinating seamlessly with each other.” Utkarsh Marwaha, Co-Founder of Mindcraft Labs
Since speech speed and temperature are configurable, Everise could configure the AI to ensure the bot was understandable across different regions. A neutral accent was used to ensure clarity and accessibility for all users. Since most caller issues were solved without needing to transfer the call, 600 man hours were saved.
"The voice and conversation type that Retell AI has is unparalleled. Whenever we play this to any of our clients and guests, people are amazed at how intelligent and human-like the technology has become." Saurabh Saurabh, SVP of Digital Transformation, Everise
Retell AI offered the power and flexibility to build sophisticated, multi-prompt agents for each individual use case. This maximized the AI’s potential to problem-solve without need for a human agent while facilitating easier QA and optimization.
“In a single prompt agent, you have to put in all your APIs outside, which means they are globally accessible in terms of scope. That increased the chance of misfire. With multi-prompt, you can actually lay down the scope of every API.” Utkarsh Marwaha, Co-Founder of Mindcraft Labs.
Everise had previously deployed voice bots using Microsoft and AWS solutions but found responses unnatural, slow, and robotic. “That was a big differentiator in us choosing Retell AI,” says Saurabh Saurabh, SVP of Digital Transformation at Everise. “You get such low latency and such consistent results all the time.”
“They have somehow cracked this on the back end,” says Utkarsh Marwaha, Co-Founder of Mindcraft Labs. “They have a good idea about what things need to be API based, what things need to be on-prem, and a turn-taking algorithm that beats the rest.”