FranceFinance and insurance2016 – ongoingBNP Paribas
Chat2Trade interpretation of trading quote requests
BNP Paribas said Chat2Trade structured around 50,000 quote requests monthly
Problem
Global Markets salespeople manually interpreted chat and email requests for quotes. They entered product details into pricing tools, sometimes taking several minutes per request.
Approach
The bank's Global Markets Data & AI lab developed Chat2Trade in 2016. The deployed system used internally fine-tuned BERT models to classify requests and extract entities. It converted the text into structured inputs for pricing tools.
Results
Takeaway
Chat2Trade automated request interpretation before pricing. The reported sub-second step did not measure the entire client transaction.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
BNP Paribas explained how Chat2Trade turned customer messages into pricing-tool inputs. It reported around 50,000 requests monthly and a request-entry step taking less than a second.
Full details are in the original sources.
