RAKT, an insurance company, implemented a chatbot to handle customer queries.
Despite its potential, the chatbot underperformed, leading to dissatisfaction.
The company identified six key issues for improvement:
latency
linguistic nuances
architecture
dataset quality
processing power
ethical challenges
Your goal is to familiarise yourself with the case study (see pdf above) in order to answer Paper 3 exam questions.
is a 500-word summary of the study to help you get started.
Once you have read the summary, follow these steps:
Answer the following 10 comprehension questions.
Compare your answers with your study group.
Continue on to read the full case study - make notes of key terms.
Attempt the exam question sample set. For questions in the set that require further reading, do your own research. This is a great way to prepare for Paper 3.
Summary Comprehension Questions
What was the primary reason RAKT implemented a chatbot?
List the six key issues that RAKT identified in their chatbot’s performance.
Why did latency become a problem for the chatbot?
What are the five stages of natural language processing (NLP) mentioned in the case study?
What is the main limitation of recurrent neural networks (RNNs), and how can it be mitigated?
Why is dataset quality important for chatbot training, and what types of biases should be avoided?
What kind of hardware is required for efficient chatbot processing and training?
List three ethical concerns related to chatbot deployment.
What are two key improvements RAKT can make to enhance the chatbot’s ability to understand context?
How does the case study highlight the importance of ethical AI deployment?