Earlier this month Yandex Data Factory, a division of the Russian search giant, unveiled ‘Extract,’ a new solution to search and track of business-critical information.
One of Extract’s first use case involves Russia’s central bank, which needs to find all cash-loan providers represented online in order to check their compliance with regulations. To address this needs, ‘Extract’ offers a custom web search service, providing a database of websites and social media pages that almost certainly belong to loan providers.
‘Extract’ also classifies these resources by their likeliness of having the required license “with a precision of approximately 71%.”
The solution is built on Yandex’s proprietary machine learning algorithms and global Internet index. It is trained specifically to find relevant websites and pages online.
Now in closed beta version for the Russian market, the service is scheduled for official launch in 2017.
“We are inviting potential B2B users to test the service now. Based on their feedback, we will make the necessary adjustments and changes to the product,” said Elena Samuylova in an exchange with East-West Digital News.
Further international roll out
Currently, the service is optimized for the Russian language, though the search in English is also possible. “Since Yandex has a global search index, we can further expand it to other languages as well. These aspects will be defined further in connection with overall plans for international launch,” Samuylova said.
Asked about competing technologies, Samuylova answered: “As far as we know, there are no direct competitors for such service. Of course, search engines are used for manual search and monitoring of relevant information – which is precisely the type of work that we aim to automate through the use of ‘Extract.’ There are also a number of niche services that solve specific tasks: for example, brand monitoring across social networks. But such services usually work with a limited number of sources, whereas ‘Extract’ has the unique advantage of using Yandex’s Internet index.”
Rather than relying on rules or keyword search, which do not always provide relevant information, ‘Extract’ leverages Yandex’s machine learning technologies. These allow finding “similar” sources, or “search by the meaning,” – in which Yandex Data Factory sees a unique feature.
Update April 2017: In April 2017 Yandex shut down the ‘Extract’ service. “We realized that customers were primarily interested in media monitoring – which as such is already offered by ‘Yandex.Mediana’,” a Yandex representative stated in a Facebook post.
Yandex Data Factory will rather focus on industrial applications, with AI-based solutions to “increase productivity, reduce costs, and improve energy efficiency.”
