Harbinger Group

Challenge
Our client provided sophisticated leadership coaching and development to global businesses. They were solely dependent on a manual process to identify patterns, analyze behaviours, and deliver executive training. This process posed various challenges as:

➢ Time-consuming data collection➢ Inconsistent and poor data quality➢ Difficulty in pattern recognition➢ Limited operationalscalability➢ Ineffective coaching planning➢ Lengthy training completion time

Our client sought to advance their leadership development planning and analysis. The idea was to deliver personalised coaching to employees through efficient content retrieval. They wanted to establish an automated solution to evaluate and classify executive behaviours, discover leadership patterns, and boost employee performance.

Solution
To overcome such critical challenges and meet essential business requirements of our client, Harbinger used RAG andagentic framework to build a robust Conversational AI platform.With RAG approach, our team connected the AI agents to anextensive repository of coaching content, including modules forskill development, leadership, and employee growth.

AI agents retrieve the most relevant and context-specificinformation dynamically from potentially hundreds of coachingmodules to access. They intelligently decide which data sourcesto use based on the user’s intent and the specific coachingcontext.

AI agents efficiently select from a wide range of sources andapplies the RAG approach to generate precise, tailoredrecommendations and insights.

As a result, HR professionals and managers can now providemore personalized coaching experiences. This enablesemployees to receive just-in-time, relevant guidance alignedwith their unique career goals and development needs.The overall system architecture integrates these componentsseamlessly, empowering the conversational AI platform todeliver a dynamic and scalable solution for coaching anddevelopment.

The system architecture discovers leadership and theircommunication patterns, analyzes their behaviors, and classifiesthem into specific categories to provide custom coaching plans. It improves coaching precision and streamline processes for planners and principal scientists.

Here are some key features of Harbinger’s Conversational AI platform backed by RAG and agentic framework:

❑ Comprehensive Data Collection and Analysis➢ Automated data gathering➢ Data reliability and logging

❑ Deep Behavioral Insights➢ Analysis of subtle behavioral nuances➢ Pattern recognition and measurable outcomes

❑ Personalized Development Plans➢ Tailored recommendations➢ Targeted leadership improvement➢ Strengths-based coaching approach

Unique Differentiators1.Intelligent AI agentic framework to dynamically assess user context.2.Contextual and relevant coaching source selection for each unique use case.3.Adaptive coaching plans that are highly personalized and responsive to individual needs.4.Scalable and extensive library of coaching modules for leadership growth and adaptability.5.Precise and just-in-time retrieval of relevant executive coaching content.6.Conversational flexibility of AI agents for optimal and personalized leadership development.

Results
Harbinger’s Conversational AI platform helped HR professionals and managers provide more personalized coaching experiences. It enabled employees to receive just-in-time, relevant guidance aligned with their unique career goals and leadership development needs. Here are some significant benefits observed by our client:

1. 50% faster discovery of relevant content for coaching planners2. 30% reduction in manual efforts for coaching plan creation3. 20-hour monthly time saving with automated content retrieval4. Increased leadership coaching efficiency and effectiveness

Explore how Harbinger leveraged RAG and agentic AI to fast-track the creation of personalized leadership coaching plans with automated data analysis.

Challenge

Our client provided training to its floating staff on standardoperating procedures, policies, and recipe preparationguidelines. However, assimilating a multitude of informationregarding the workflow was challenging for employees.During peak hours, their support center was overwhelmed withmultiple calls, leading to delays in resolving employee queries.This led to a recurring issue of high employee turnover.On the other hand, many of the franchisee employees wereMexican and preferred query responses in Spanish. Thefranchisee wanted to introduce a customized functionality thatallows employees to receive responses in Spanish. Theirobjective was to ensure ease of work, employee satisfaction,and productivity.As such, the company sought to accelerate their employeeperformance support process. They wanted to generatemultilingual and more accurate query responses for theiremployees. Their aim was to create a standardized responsegeneration procedure as per the proprietary documents.Some of the other key business requirements of our client wereas follows:➢ Automate query resolution mechanism to provide 24/7employee performance support.➢ Keep a track of the types of queries, responses generated,and employee engagement.➢ Build a hands-free system to ask questions and receiveimmediate verbal responses.➢ Avoid the involvement of high-priced support experts andtranslation professionals.➢ Handle a high volume of queries with utmost precision andminimal human intervention.

Solution

Development of AI-Based Voice-Enabled Chatbot

❑ Engineering and Training of AI Model

Harbinger developed and trained an AI model based on our

client documentation consisting detailed instructions to carry

out assigned tasks. This AI model was integrated into a chatbot

accessible to franchisee employees at the point of sale or within

the company system where they were logged in.
 

❑ Fine-Tuning of AI Chatbot

Harbinger integrated the AI-based chatbot with ‘thumbs up’ and

‘thumbs down’ feedback options. This helped identify whether

the responses provided to the employees by the AI chatbot

were helpful and accurate.

Harbinger gathered this feedback and analyzed the instances of

thumbs down. The AI chatbot was fed with the right answers

and trained on certain instances to provide employees with

relevant responses.

❑ Multilingual Response Generation

Harbinger integrated the AI chatbot with Spanish language

support. The integration eliminated language-specific document

constraints to enable cross-language interactions. Employees

could ask questions in Spanish and receive responses based on

an English knowledge base, with the final output in Spanish.

In the same way, the AI chatbot accepts queries and provides

accurate responses in various other languages.

❑ Implementation of Dashboard and Analytics

Harbinger integrated the AI chatbot with a dashboard backed by

sophisticated reporting and logging mechanisms. It involved

analyzing employee interactions and system behavior.

The information was used to create reports enclosing the list of

queries requested and responded to and from different

locations and devices.

❑ Integration of Alexa Skills and Conversational AI

Harbinger integrated the AI chatbot with Alexa Skills. The

integration helped our client’s floating staff receive immediate

answers to their queries in natural language.

Through Alexa Skills implementation into the AI chatbot, our

team created an AI-based voice-enabled system to strengthen

employee performance support.

This phase focused on the following key aspects:

➢ Comprehensive analysis of knowledge base for employee

support.

➢ Utilization of Alexa Skills for instant verbal response to

queries.

➢ High natural language accuracy for specific terms and

instructions.

➢ Conversational AI interface to reduce employee time and

effort.

Results

Unique Differentiators:1. Precise AI model development, training, and fine-tuning to generate error-free employee query responses.2. The chatbot is scalable enough to handle a high volume of queries, ensuring consistent performance even during peak times.3. A dashboard with robust reporting tools analyzed employee interactions, helping refine and optimize the system over time.4. The chatbot employed advanced NLP techniques to understand and respond to natural language inputs with high accuracy.5. The chatbot automates routine query resolution to reduce the need for high-priced employee support experts.6. The chatbot delivers accurate responses as per client database, keeping it secure and isolated from unwanted or external sources.Harbinger’s AI-based chatbot helped the company accelerate their employee performance support process. The chatbot eliminated language constraints by accepting and providing accurate responses in multiple languages. Here are the other significant benefits experienced by our client are as follows:

1. Enhanced query resolution with 92.5% success rate2. More than 90% accuracy of employee query responses3. Improved employee performance and engagement4. Automated, faster, and verbal response generation5. Increased time savings and reduced training costsExplore how Harbinger developed an AI-powered voice-enabled chatbot to automate and transform employee performance support for the world’s largest hot dog chain.