|
Getting your Trinity Audio player ready...
|
The following article describes a solution we built for one of our clients – an AI-based chatbot for fast, accurate, and high-quality customer service, with significant cost savings and quick return on investment.
Background on the client – Timewatch Company:
Timewatch provides the smartest tools for time and employee attendance management through online time clocks and supports thousands of customers.
The company’s challenge:
To provide support to thousands of customers in a fast, efficient, and high-quality manner, so they can receive quick and accurate responses 24/7 without needing to wait for a human representative or necessarily needing to call for support.
Therefore, the company contacted us to analyze the problem, find an appropriate solution, and implement it.
The shared challenges we faced:
- Working with the organization’s existing tools (in this case, Zoho Live Chat and Zoho Knowledgebase)
- Building a fast and accurate indexing capability that could also work automatically and continuously (once a day, to add changes to the Knowledgebase)
- Working with AI tool APIs in an agnostic way that would allow easy switching between them (to switch transparently for the user if a certain tool isn’t accurate enough or if its terms of use are changed by the manufacturer)
- Building a quality solution that would also support articles in Hebrew
- Connecting all the components into a quality, reliable, and fast solution.
The solution we proposed:
Through collaborative work, we defined how end users (customers) would work through Chat and what would run behind the scenes.
The solution we built is based on AI (using GPT and Claude and working with their APIs), RAG and a Vector Database, which learned the company’s internal Knowledge base (and all relevant accumulated knowledge) so it could provide accurate answers from the precise context related to the question the customer is asking.
This way, customers can effectively ask questions in free text via the Chat and receive immediate answers.
This is the Chat interface that was added to the customer user interface:
The solution we built also included infrastructure aspects (DevOps), automation, and Scale (in which our company is also very experienced), to support ongoing updates, support for many simultaneous users, ability to separate development and production environments, recovery and resilience capabilities, and more.
This way, all parties benefit:
A. Customers receive immediate, professional, and accurate support 24/7.
B. Timewatch, which provides the service to customers, can provide faster, more accurate support, reduce response times while achieving significant cost savings, and maintain a high reputation.
“Using Chat based on AI reduced the load on our customer service and improved response times for our customers and their user experience“, summarized Ronen Porat, CEO of TimeWatch.
