Basis Technology Powers Streamfully, Boosting Web User Engagement by Deepening PersonalizationJanuary 13, 2015
Innovative Conversation-Like Interaction Surfaces User-Relevant Content in Real Time
CAMBRIDGE, Mass., Jan. 13, 2015 — Streamfully, from New York-based General Cybernetics, Inc., is the latest innovative startup taking advantage of Basis Technology’s program that makes professional-grade text analytics available to early-stage, high-impact startups.
Streamfully increases visitor engagement and interest in a website through a highly personalized conversation model. Streamfully’s algorithms use noun phrase extraction and entity extraction from Basis Technology’s natural language processing platform, Rosette, to feed a proprietary algorithm that ranks choice snippets of content, whether from a hovered-over link or from across the web. Responding to the user’s interactions, Streamfully’s recommendations become more and more refined.
“Related Links” that are commonly displayed on publisher sites are not derived from the user’s interests, or, at best, are an extrapolation based on a statistical assessment of “users-like-you.” In contrast, snippets previewed by Streamfully respond to each user’s individual context and prior browsing. Therefore, users gain more value from content uniquely tailored to their interests. At the same time, the content provider exposes more site content than the users might find on their own. Furthermore, this high level of personalization enables higher quality and better-targeted ads and sponsored content.
“Streamfully is based on the idea that the most effective user interface is conversation. The immediacy of back-and-forth on a topic gets to the point faster and more accurately than any number of keyword searches,” Paul Pangaro, President of General Cybernetics, said. “Streamfully gives users interacting with web content the benefits of a conversation.”
How Streamfully Works
Streamfully first analyzes pages for thematic segments, which are processed by Basis Technology’s Rosette Morphology to extract key phrases and Rosette Entity Extraction to find people, places, and organizations. Streamfully builds on Rosette’s analysis to semantically quantify differences between texts that seem to have similar content. For example, imagine three pages containing “Julia Child” and “chicken.” One may be about the supposed “dropped chicken” episode; the second, a Julia Child roasted chicken recipe; and the third, a Julia Child letter that discusses “buttery, French chickens.” Streamfully’s algorithms calculate which page is most likely to be interesting to the user, based on the user’s context and prior interactions. It then suggests that page to the user—mimicking good conversation.
“Rosette is absolutely crucial to what we are doing. We need an extremely reliable and efficient way of pulling out key phrases from unstructured text, and Basis Technology seems to have nailed this,” Paul Pangaro, President of General Cybernetics, said. “Rosette gave us the raw power of extraction and recognition, and the flexibility to run in our stacks without modifications, whether in our in-house development servers or at cloud-scale.
“We looked at open-source options and none were even close. Having access to Rosette through the startup program meant we didn’t have to spend time implementing our own or working around the limitations of other commercial systems that had more limited features. With the quality of results Rosette delivers, we can focus on this very hairy stuff that makes Streamfully unique.”
“Streamfully is exactly the type of innovative, high-impact idea that our startup program seeks to champion,” Carl Hoffman, CEO of Basis Technology, said. “The idea that the computer can process volumes of information that are beyond human scale, and then with a little conversation with a human, find exactly what is sought, is another great idea in the human/computer collaboration future that we see.”
About General Cybernetics, Inc.
General Cybernetics is changing how users discover and explore digital content through deep personalization and real-time hyperlinking. GCI’s user interfaces and services platform deliver content that is always relevant and novel by building a fine-grained “interest history” for each user. GCI’s first product is Streamfully, a browser extension that looks ahead into web links and creates personalized content previews.
About Basis Technology
Basis Technology provides natural language processing (NLP) solutions for extracting meaningful intelligence from unstructured text. The Startup Program makes our Rosette linguistics platform available to high-impact, early-stage companies at an entry-level price. By starting with professional-grade text analytics, these innovators can focus on realizing their core technology instead of wasting valuable time re-inventing text analytics or adapting tools that may not scale or have the accuracy required by a production system.