The Business Challenge
Dedicated to making better tools and making a better world, our client’s mission is to build trust with those who build and shape the world around us. Crucial to building trust was consistently delivering experiences across touchpoints that were intuitive, easy and satisfying.
In tools industry
Business vertical
Global presence
With a focus on increasing positive buyer and customer sentiment, the manufacturer identified two areas with the opportunity for improvement: virtual shopping and issue resolution.
For the former, the client aimed to create a virtual shopping assistant that would engage customers in interactive conversations — offering personalized product recommendations based on the user’s requirements to improve the shopping experience and boost sales. However, crafting a customer-facing system like the one the client wanted had the added challenge of requiring robust guardrails to ensure interactions are secure, accurate and focused.
And for servicing its customers, the client was keen to elevate its efficiency in discovering knowledge through KeDB articles and resolution notes from previously resolved tickets. Manually, it can be very challenging to search through the landscape of KeDB articles, resolution notes, ticket dumps and more. Congruently, creating new KeDB articles is time consuming for already time-sparse employees.
The manufacturer needed a partner that could deliver the automated solutions to meet its needs at both ends of the customer journey spectrum.
- Boost sales by helping customers to locate their desired products quickly
- Compare products and recommend the one that best fits customer needs
- Enhance customer satisfaction by accurately identifying and addressing their needs
- Apply generative AI search and application support with natural language query-based response
- Improve MMTR / time to resolution
The AI Transformation
Working in collaboration, HCLTech delivered solutions that propel its ecommerce and issue resolution into the future.
Utilizing the latest advancement in generative AI (particularly multi-modality), HCLTech built a custom assistant that provides personalized guidance to shoppers. By leveraging the latest GPT-4 Vision model, the solution ensures they find exactly what they are looking for. A guardrails security layer ensures conversations stay on topic by restricting them to product recommendations only. The generative AI backbone supports conversations based on images the customer uploads — using learned data to deliver top-notch product recommendations to boost sales and enhance the overall shopping journey.
Equally as important to creating satisfying virtual shopping experiences is resolving issues quickly. To address this need, HCLTech built a robust generative AI solution using Microsoft Azure Stack. With Microsoft Azure AI Document Intelligence, we automated data processing using AI and OCR to enable swift information extraction. We delivered a rich and intelligent search experience through Microsoft Azure AI Search. And utilizing Microsoft Azure OpenAI’s industry-leading coding and language AI models with Prompt Flow’s streamlined engineering capabilities for back-end build, paired with a front-end interface built using ReactJS, HCLTech was able to expedite the deployment of the solution.
The deployed solution restructures the KeDB articles, runbooks and ticket dump to improve data quality — and new articles can be generated using historical ticket data and uploaded to the repository to improve quality of documentation. The implementation of a feedback loop enables the solution and services to help optimize those efforts, recommending opportunity for automation, problem management and shift left.
With the added integration of the cloud-based service Microsoft Azure AI Translator, questions and responses can be made multilingual, aiding in further improved efficiencies.
The Business Results
From assisting ecommerce shopping experiences to resolving customer tickets efficiently, the solution is realizing tangible results for the client:
Questions answered with a high degree of accuracy and relevancy
L1 Teams of 2 Towers are using the solution
Improvement in MTTR is being realized
