A leading digital lender with $1B+ AUM slashes median customer support wait time from 4 hours to 5 minutes.

A leading digital lender with $1B+ AUM slashes median customer support wait time from 4 hours to 5 minutes.

Frontrunner lending platform with over 1M customers responds to over 3K emails each day with AI powered responses using previous customer interactions, company's knowledge base and a secure API integration over FreshDesk.


Leading Digital Lender
Bangalore, India





CX Agent Assist

Lower average wait time for email queries
Contact centre productivity improvement
Customers addressed daily over emails
  • 5 hr → 2 min
    Draft response generation time
  • 2500+
    Agents hours saved per month
  • Knowledge

    20,000+ previous conversations converted to knowledge, website content pages, user profile from CRM

  • Skills

    Draft ready-to-send email response with placeholders for customer specific variables

  • Channels



The client is a Series D funded financial services platform that offers retail and business loans with over $1B annual disbursements, $40M in annualised profit target and over 7 million customers.


The client, a regulated fintech company, was seeking to reduce its operational costs by improving its support contact centre. They receive over 10,000 tickets per day, with half coming through emails and the rest through phone calls and in-app chat. However, customers have to wait an average of 4-5 hours for a satisfactory response. This is mainly due to the pre-opening hours ticket queuing and ticket assignment delays. A relatively smaller time is taken by support agents to draft email responses.

One of the challenges faced by the client was the lack of utilisation of company wide tribal knowledge and the company's knowledge base when drafting email responses. With thousands of tickets flowing each hour, the manner in which the human agents responded to customers with different account profiles, default statuses, credit scores, etc. was a treasure trove. 

Moreover, there were inconsistencies in the quality of responses as it depended on the support agent's tenure and knowledge. New agents also took a significant amount of time (3-6 months) to become as effective as more experienced agents.

The goal for the client is to reduce customer wait time by 50% while improving the quality and personalization of email responses. They also aim to decrease dependency on agent tenure and training levels while bringing more standardisation to the responses. Additionally, they hoped to increase agent productivity by 30-40% and reduce ramp-up time from 6 months to just 1 month.


The client and Alltius joined hands to slash the overall customer wait time for email responses. The team from Alltius, consisting of engineers and product managers, visited the Contact Centre to observe support agents as they responded to emails. They also interviewed the contact centre managers to gather their hypotheses on how to decrease customer wait time.

The proposed solution involves implementing a highly secure AI assistant powered by the Alltius platform. This assistant would be coached using a vast  dataset consisting of resolved customer conversations, web documentation, and support agent handbooks. Through this training, the assistant would not only acquire knowledge but also familiarise itself with the company's typical response style and objection handling strategies. 

Leveraging this training, the assistant would generate a personalised initial draft email response that includes relevant placeholders for customer-specific details. These placeholders would then be replaced with specific values from within the organisation, ensuring that no sensitive information is shared or exposed to public LLMs. The agents could then accept or editorialise the suggested email response and finally shoot an email response off to the customer.

The client successfully integrated this assistant with their FreshDesk agent workbench using Alltius’ asynchronous and polled API channels. The payload included customer details and their email query, while the response consisted of a well-crafted email generated from a synthesis of previous email exchanges, knowledge base information, and placeholder variables. This integration was tested on over a hundred sample tickets that were reviewed and approved by the business and contact centre operations. Upon clearing the quality thresholds, the agent was cleared for use by agents in real time.


The assistant has been upgraded and is now generating 3,000-5,000 email responses per day. This has led to a 50% decrease in the average waiting time for customers, which now stands at around 2-3 hours. 

Instead of spending a lot of time crafting responses, agents now focus on validating them. The majority of the necessary information for personalization is already filled in and drafted beforehand. Thanks to Alltius' asynchronous responses, the wait time before the contact centre opening hours and the time drafting an email is significantly reduced to only a few seconds. Consequently, the assignment queue is much shorter, resulting in a decrease in overall customer wait time.

Way forward

The main objective of Team Alltius is to enhance the quality of draft emails to a level where the additional effort put in does not justify the marginal improvement in quality. Next, our aim is to expand our coverage to handle all 5000+ customer emails received daily. Additionally, we will be conducting pilots to address in-app chat scenarios. Furthermore, we will be implementing ticket assignment and labelling use cases to reduce customer wait times.

Happy customers. Quickest time to value.

Matchbook AI

Matchbook's data intelligence platform now offers Alltius powered AI assistant to its users from marquee Fortune 500 brands, where they can not only get instant help from across their documents but also raise tickets from within the bot.

Read Story

Hiro System helps thousands of crypto users by helping them locate precise help and code snippets in seconds from an abundant overwhelming pool of information and resources.

Read Story
Leading Asian Bank

The large multinational bank can now give instant recommendations to its 12 million customer base on its card products and near instant insights to its large equity analyst team with high quality interpretation of chart and table filled research reports

Read Story

Your journey starts here

Let us personally show you how Alltius can help your product.