For this project, Mastercard’s R&D Labs needed unique, multi-lingual text data that covered 20 designated payment scenarios in both English and Spanish, and they needed it fast.
Obtaining comprehensive data under highly specific parameters is a multi-step process that can prove time-consuming without experts involved. We worked side-by-side with Mastercard to ensure maximum efficiency and quality at every stage, from assisting with scenario development, to drafting a comprehensive set of prompts, to collecting, validating and annotating all collected text.
Start to finish, this entire process needed to be completed in less than three weeks.
Our well-tested workflows and curated Neevo workforce were designed to handle exactly this sort of project and timeline. We sent original prompts covering all 20 payment scenarios out to our Neevo contributors, who produced 2,000 variants for each language in response. We had each of those variants validated and corrected by separate crowd members, and for the 475 that were questions, we collected 1,500 answers and had each one ranked for relevancy three times.
Finally, we had every variant annotated for entities by three separate Neevo members. In the end, we collected over 6,000 entities split between English and Spanish. We took out the repeats and delivered over 1,000 entities in each language to Mastercard.
When it comes to AI, people tend to say data is “king.” That’s not quite the whole story. Companies like Mastercard understand that precision within data is the throne, crown, and the keys to the castle. In partnering with DefinedCrowd, the R&D Lab benefited from exactly the kind of precise, industry-specific training data that will fuel another decade of innovative leadership, delivered with: