AI • June 27, 2022

Building Smart and 24X7 Service-ready Help Desks With AI Chatbots

Enterprise contact centers and service desks face high operating expenses primarily due to enormous staff  and training costs. Chatbots are a convenient option for them to support their help desk and other customer-facing teams. They can enable end-users with self-serve features, reducing help desk workloads and improving employee productivity. As customers demand faster resolutions, self-serve options become a necessity.

Different Chatbots for 24x7 Ready Service Desk

Advances in AI, particularly Natural Language Processing (NLP), have enhanced chat and voice bots convertibility, query resolution capabilities, and user experience. Today, chatbots have transitioned to intelligent, conversational-driven' Virtual Assistants' from command-driven apps which are adept at determining context and user intent. 

Based on the degree of intelligence and use case, chatbots' capabilities vary from answering questions to servicing an employee. Chatbots can be broadly classified as:

  1. FAQ chatbots that understand questions and provide answers.
  2. Virtual Assistants are integrated with enterprise systems to perform auxiliary tasks such as booking a hotel, lodging a complaint or scheduling an appointment.
  3. Virtual Agents chatbots that are highly intelligent and can handle complex dialogues, processes and security protocols .

Building the chatbot

Before developing chatbots, enterprises need to define the bot's functions. Architecting a chatbot requires an in-depth understanding of 

  • Customer's IT landscape, 
  • Bot components, 
  • Frameworks needed, and
  • Accelerators/platforms to be used. 

The chatbot must be human-like to drive effective conversations and improve user experience. An industry-specific chatbot is an ensemble of multiple models working together to understand the language and context. 

Chatbots have to do all of the above without giving themselves away to a troubled employee/customer reaching out to the help desk. This is a tough task but the right implementation partner can make it happen. Let's understand it better with an implementation by Quantiphi for one of the largest energy holding companies in the United States.

Attributes of Successful Chatbot Implementation

One of our customers, a North Carolina-based Electric Utilities and Infrastructure business company that serves over 7 million customers, wanted to enhance their support user experience and the associated internal processes while reducing support headcount and redirecting its technology team.

The Challenge

Before chatbot implementation and help desk process modernization, the customer was dependent on the support team to manually resolve internal queries. It was a time-consuming process where all employees had to call a helpline number for support-related tasks such as resetting the password and opening an account. The agents then resolved the query on the call itself. For complex requests, they manually updated support forms with information gathered from the user and created a ticket.

Extensive manual work and voluminous support issues led to errors and affected operational efficiency. There was also a lack of correlation for real-time updates between various client internal systems, including their Azure environment, BMC Remedy and LDAP. Changing the existing process was tedious due to security concerns, with limited APIs available, and even then, it needed custom efforts and third-party services. 

Solution Implementation

Quantiphi created a natural-language-powered conversational virtual assistant with text and voice capabilities using various AWS services to support the client's Help Desk team. 

  1. The solution leveraged AWS Lex to develop a Voicebot and a Chatbot that can manage several conversational flows. This eliminated the need to involve customer service representatives and engineers to address and resolve support issues.
  2.  Amazon API gateway's ability to handle web connections in real-time ensured that the users receive prompt responses.
  3.  The strategic use of AWS Lambda and Cloudfront ensures a fully managed, serverless, and scalable solution. It also ensures confidentiality as sensitive data is secured with Cloudwatch, Route 53, WAF, KMS, IAM, Config, CloudTrail, Secrets Manager, etc. and stored using AWS DynamoDB. 
  4. Every interaction became a data point, which was subsequently analyzed using AWS Glue and Kinesis, and then displayed against the client's KPIs to discover leakages and vulnerabilities in the system.
  5. Account management was enabled for the client's users while adhering to industry best practices in cyber security. Each user's identity was verified against the existing database using one of two custom identity verification protocols: Manager's Approval or Self-Verification. The solution was developed to send a verification email to the user's reporting manager, who confirmed the user's identification, for Manager's Approval. If the user chose self-verification, a verification call to the user's MFA registered phone number was initiated using AWS Connect to authenticate the user's identity. Once validated, users could administer their own accounts, reducing the need for manual intervention in the support process. Also, the solution's linkage with the client's ITSM interface allowed them to generate, track, comment on, and escalate tickets. The current knowledge base was also linked, allowing them to search for articles on often encountered problems.
  6. Customer satisfaction was ensured through AWS Connect's call conferencing features to perform a live agent transfer on use cases outside the solution's scope. By providing business intelligence capabilities, performance indicators such as the total number of sessions, average session time, live agent transfers, and abandoned sessions were tracked. These data points helped in  identifying the pain points and further iterating on the solution.


The solution enabled the client to self-service its employees for support-related queries 24/7. It was scalable, consistent and cost-effective and reduced the need for support staff to address questions. Approximately, 33000 support queries are projected to be resolved by the bot with minimal intervention annually.

  • The developed solution created tickets on behalf of users, allowing them to check their current ticket status, and track ticket history without any agent.
  • It also escalated tickets based on pre-defined business rules.
  • The in-built business intelligence of the solution quickly identified gaps in their knowledge base and any recurring issues for the defined KPIs.

We continue to build further integrations and develop extensions with their other internal applications. We are also enhancing the existing solution to address additional use-cases for the Enterprise Helpdesk teams.

Quantiphi's Contact Center Transformation practice transforms the client's contact center landscape into a scalable cloud-based enterprise through the power of AI-driven Chatbots and Voicebots. Quantiphi's solutions add value to the existing workforce by addressing multiple use-cases whilst seamlessly integrating with existing systems. 

Get in touch with us to transform your helpdesk and maximize value with customer interactions with conversational AI.

Contributed By: Varun Tewari, Suraj Vantigodi

Written by

Rohith Krishnan

Thank you for reaching out to us!

Our experts will be in touch with you shortly.

In the meantime, explore our insightful blogs and case studies.

Something went wrong!

Please try it again.