Tell us about your project!
Finance
Collected data improved accuracy of customer intent understanding by 8%
Training data for AI and machine learning in finance
About case:
About case:
Industry and use case
Finance, consumer intent understanding
DATA
50,000 sentences
Project duration
21 days
Industry and use case
Finance, consumer intent understanding
DATA
50,000 sentences
Project duration
21 days
Challenge:
Challenge:
The client was developing a chatbot in the financial sector and wanted to improve the accuracy of consumer intent understanding. They contacted Training Data solution to enlarge their current dataset so that each intent from the dataset could be rephrased in multiple different ways. They were looking for a team familiar with the financial sector terminology.
Chatbot message «How do I view the balance of my debit card on my mobile banking app?»
Chatbot message «How do I view the balance of my debit card on my mobile banking app?»
- Where do I find my balance?
- How much money is in my account?
- Where can I find my money within the app?
- How much money do I have left?
- How do I check the cash balance of my account?
- Why can’t I see the balance for my card in the app?
- Is my account balance on the app?
- How much money do I currently have?
- Can you check the balance on my card?
- How do I request my balance?
Training Data Solution assembled a 15-person team with previous backgrounds in finance. Each intent from the initial dataset was then enriched with ten different ways to express it. As a result, the training data increased ten folds.
Outcomes:
Outcomes:
More requests are solved without human intervention
Collected data improved the accuracy of intent understanding by 8%
Feedback
The intent text collection was done quickly, and the team at Training Data Solutions implemented excellent quality control within our project. Training Data Solutions is a very trustworthy partner!
Tikhon E. Project Manager