IBM and Colombia announced a strategic partnership to leverage IBM Watson to augment regional language content promotion, implement state of the art cross-language recommendation systems over multiple media contents- text articles, audio, video etc.
As one of the largest publisher-owned ad network platforms in APAC, Colombia was looking for a technology partner who can help address the growing need for varied content formats for targeting their publishers as well as readers. Regional language content being the key, the ability to promote content across language was a key necessity to stay ahead in a competitive environment.
Colombia leveraged IBM Watson’s Natural Language Understanding service to enrich content with additional features such as named entities, concepts and categories. These raw feature enrichments, significantly improved the platform’s up-stream real-time recommendation algorithms.
Swapnil Shrivastav, CTO, Colombia said, “We chose to work with IBM Watson as we found its Natural Language Understanding capabilities to be the best fit for our requirement. By leveraging IBM Watson, we have built a home grown proprietary technology stack, tailor-made for our use cases and to help further strengthen our publisher’s digital presence beyond regular social channels. Furthermore, our reader’s interaction with the platform in terms of retention, viewership, and readership has increased. We have witnessed a three-fold increase in contextual recommendations with a focus on cross lingual endorsements.”
Amit Kumar, Cloud Leader, IBM India/South Asia said, “We are excited to be the strategic technology partner for Colombia in helping them offer quality, relevant, multi-lingual, and real-time content to their publishers and end readers via platform of their choice. IBM Watson’s Natural Language Understanding service will help Colombia get insights from their data to build their multi-lingual platform as they plan to grow their vernacular content synthesis, information transmission across audio, video and concise media snippets format.”