Project snapshot
Our long-term client, seeking to streamline and consolidate their multi-channel support system, embarked on a project to create a unified AI-powered support assistant that helps employees resolve work-related issues and reduces the support team’s workload. Intellias initially joined the project to augment the backend team, but the collaboration quickly grew, with our engineers forming the core team for solution development.
Together with the client’s software development team, we built a comprehensive support platform that improves workplace experience and provides high-quality assistance to thousands of company employees.
Our client, a multinational telecommunications provider, uses advanced technology to improve communication quality for customers in more than 80 countries around the globe. Our collaboration spans many years and includes a number of successful joint projects, such as an employee workplace management solution and a customer self-service portal.
Business challenge
Focused on increasing the speed and quality of employee support service, our client implemented several chatbots to help users resolve issues and navigate the company’s extensive knowledge base more effectively. The telecom provider was running as many as 14 different chatbots serving various departments, which presented certain challenges for employees who had to use multiple support channels to resolve issues.
To implement a single point of contact for all matters related to user service and support, our client decided to unify the existing support channels into one platform leveraging artificial intelligence to help users find necessary information. The goal of the project was to combine knowledge base search, ticket generation, and live agent support into a cohesive solution providing a unified user experience.
To ensure the best support quality, smooth user experiences, and optimized costs, the company opted for an AI-powered support assistant using a large language model (LLM). The implementation turned into a very complex project with a distributed team focused on developing the three major components of the solution:
- Frontend: Represents the user interface and related interactions
- Backend: Houses the infrastructure and enables the business logic of the solution
- Bot platform: Integrates multiple front-end interfaces and allows for potential integration with external applications.
Intellias first joined the project to provide backend development expertise, addressing a skill gap in the client’s team to ensure timely completion of the project. At a later stage, our experts participated in the development of all three components, driving core processes and providing input for major decision-making.
Solution
In collaboration with our client’s team, we developed an intelligent assistant based on conversational AI. The solution supports intuitive natural language interactions and simplifies the information search in the company’s knowledge base.
Responding to support requests, the LLM finds the best match between the user’s query and the knowledge base, selecting the most relevant articles and generating a user-friendly response. When building a response, the tool identifies information directly related to the query and structures it to create a human-like conversation, enhancing overall support experience. The response also includes relevant knowledge base links helping the user to access full information on their query. Upon the user’s request, the assistant can connect the user to a human agent or initiate the generation of a support ticket.
The LLM uses data contextualization to maximize the relevance of responses to user queries. The tool uses the same context to route the user to the appropriate support department and generate a support ticket form. Data contextualization also enables follow-up queries to fine-tune or expand the initial response.
The solution also provides comprehensive analytics on LLM usage, including response times, query types, and types of the model’s responses (plain text, knowledge base search, ticket generation, etc.)
Our client’s future plans include integrations with various internal systems that support the company’s infrastructure. This will allow the telecom business to create an omnichannel support solution unifying all existing points of contact into a single platform. Also, the LLM, which provides responses to queries, will be fine-tuned with the client’s corporate data to improve response relevance.
Business outcome
Intellias enhanced the client’s project with versatile expertise in multiple tools and technologies. Our knowledge and experience of using cloud services such as AWS and Microsoft Azure, OpenAI, and various development frameworks, made a notable contribution to the project’s progress and successful completion.
With Intellias participating in the creation of all solution components, the software was delivered according to the original schedule, meeting the estimated KPIs. The approaches suggested by our front-end experts allowed for a more organic and structured development process, resulting in an intuitive and user-friendly product.
The implementation of an intelligent support assistant has opened the opportunity to unify multiple conversational support channels into a single platform integrated with the business infrastructure of the telecom provider. This solution optimizes support flows for both users and support teams, centralizing issue resolution and information search.
At the same time, the solution’s built-in analytical capabilities allow the company to track the performance of the support system, follow user journeys, and identify areas for improvement.