Data is often said to be the “new currency,” as it can serve significant value when companies unlock the arsenal of qualitative and quantitative information that they generate. But sifting through the noise and making sense of unstructured data is a challenge, which is where automation and artificial intelligence could help.
Against that backdrop, Cambridge, Massachusetts-based Gamalon is working on a machine learning system capable of natural language processing (NLP) to manage billions of inbound messages, providing individual responses and communicating ideas and sentiments back to humans.
The company also announced today that it has raised $20 million in a series A round of funding led by Intel Capital, with participation from .406 Ventures, Omidyar Technology Ventures, Boston Seed Capital, Felicis Ventures, and Rivas Capital.
Founded in 2013, Gamalon had raised around $12 million in funding before now. With another $20 million in the bank, the company said that it will invest in R&D and expand its sales and marketing efforts to the enterprise.
The crux of the problem that Gamalon is looking to fix is this: Much of the data gathered by companies doesn’t live in neat, numbered columns within spreadsheets. There are email conversations, surveys, feedback forms, phone transcripts, product reviews, and so on, which often exist in such vast quantities that it’s difficult for even an army of trained humans to parse and transform it into meaningful, actionable insights. Plus, language is complex and nuanced — a handful of words can express a ton of ideas.
There are many bot-based customer service tools out there already that strive to understand customer sentiment and questions, but what Gamalon is aiming for is to move beyond rules-based machine learning systems and deep learning techniques that rely on constant training, and into the realms of teaching computers “ideas.” What we’re talking about here is unsupervised learning.
The company’s so-called “idea learning” platform is all about serving companies with “accurate, editable, and explainable” processing capabilities for inbound customer messages and other forms of unstructured data. The company said that its platform can now process natural language and then explain the ideas. “We’re creating an AI model by essentially just talking to it — and Gamalon is figuring out what ideas, and the structure of those ideas, are present, and then automatically creating a model that can then be used on new utterances and messages,” a spokesperson told VentureBeat.
“Gamalon’s mission is to accelerate human understanding by combining human and machine learning,” noted founder and CEO Ben Vigoda. “It’s easy enough for a business analyst to get started providing return on investment right away. When Gamalon’s Idea Learning technology reads large amounts of text, and forms ideas, the AI becomes an extension of you — allowing you to read and respond to huge volumes of messages.”
Part of this update includes Idea Studio, a new interface for creating and managing ideas, whether they are generated by machines or humans. Before now, Gamalon customers had no direct interface to edit these “idea trees,” whereas now machine learning non-specialists can fine-tune and improve a system’s accuracy by answering clarifying questions when the system doesn’t understand a particular response from an inbound customer message.
The idea is that humans will not need to build these idea trees, but can merely serve as collaborators for a system that will improve constantly in terms of its understanding of human language.
“The Gamalon system will enable enterprises to improve customer experience and retention, and drive revenue by allowing enterprises to truly understand what their customers are saying and to take strategic action, as well as to immediately reply one-to-one,” Vigoda added.