BBVA has recently tested an AI language model by OpenAI to see if the technology generates useful summaries of conversations between customers and bank agents to save time for customer-facing teams.
The tests, which produced “encouraging” results, were run on ten unedited Spanish conversations about banking extracted from publicly available online forums.
The financial services company tested the technology on customers without any previous customisation, re-training, or transfer learning, and asked the tech to generate summaries that would be useful to bank agents who join the conversation.
The Spanish business said that while the accuracy of the test on just ten conversations “can’t be very high,” the results obtained were “encouraging and suggest further exploration of the model’s value for the targeted tasks.”
Recent Stories