Customer Relationship
❯
Self-Service Customer Support
❯Better human-computer interaction with advanced language models
AI virtual assistant for customer support and service
For:
Customer service department
Scope:
Customer support service, product and service consulting.
Limitations - support for dialogs exclusively within MTS products.
Target audience - b2b, b2c clients of MTS Russia.
Goal:
Other
Next Level Customer Services: Email automation with natural language processing
For:
Customer support department
Goal:
Improved Employee Efficiency
Better human-computer interaction with advanced language models
For:
End users
Scope:
Human-computer interaction
Goal:
Other
How AI/ML Improves its Underwriting & Supports for Commercial Insurers with Their Agents
For:
- Insurer agents & brokers
- Carrier agents
Goal:
Improved Employee Efficiency
Better human-computer interaction with advanced language models
For:
End usersGoal:
OtherProblem addressed
Improve the quality of human-computer interaction.
Scope of use case
Human-computer interaction
Description
Natural language processing (NLP) technologies are crucial
for interaction of social systems and artificial intelligence
algorithms. AI models used in NLP are typically trained on
large amounts of text (corpora) in order to make them
learn language in general (language models). Then these
models are fine-tuned to particular down-stream tasks like
question answering, paraphrasing, fake news detection, etc.
High quality models require large amounts of data to train
and thus large computer systems are needed. Additionally,
we employ neural architecture search and automated
hyperparameter optimization techniques to derive better
models. Finally, we built a rigorous evaluation framework to
explore how model architectures and hyperparameters,
including source corpora and pre-processing methods, affect
performance of the models on each of the down-stream
tasks.
Conversational agents
AI: Act