Businesses generate a lot of dark data — unstructured data in the form of customer communications, employee emails, presentations, manuals, maintenance and research notes, etc. Enterprise-relevant data is also present on the Web in the form of blogs, social media discussions, news, etc. The heterogeneous, fragmented, and complex nature of dark data poses significant challenges in harnessing it for improved decision making and enterprises across the world struggle to make sense of it. However, recent advances in Knowledge Graphs and artificial intelligence have created lots of opportunities to tackle this critical but challenging problem.
Founded in late 2017 by Partha Talukdar, a faculty member in the Department of Computational and Data Sciences (CDS) at the Indian Institute of Science (IISc), Bangalore, KENOME helps enterprises unlock value from dark and unstructured data.
KENOME has been operational since Jan 2018, and the journey has been exciting for the founders since then for the company. “We have seen firsthand how our products can have a real-world business impact and that has been very satisfactory,” said Partha. “The traction we have received in the first year itself is extremely encouraging and we are just getting started.”
Interestingly, KENOME is bootstrapped and revenue positive startup and it believes that the company has been fortunate right from the very beginning to have paying customers. “Our pilot engagements have gone well, and we have been able to convert those into recurring revenue contracts,” said Partha. Talking about the verticals it serves, KENOME at present is active in the areas of retail and B2B commerce, finance, and consumer electronics.
How KENOME Is Using AI To Make Sense Of Dark Data
KENOME is a core AI startup. It has seven members out of which two have PhDs in AI, and together the team brings in over 20 years of relevant research experience. “AI is an enabler in everything we do,” said Partha.
Having years of world-class research and industrial experience, the company has developed methods to help unlock the true value of dark data using Knowledge Graphs. Knowledge Graphs are entity-relationship graphs which companies such as Google are using to improve their AI and user experience. KENOME democratises Knowledge Graphs for the Enterprise.
Today, the company can perform continual learning with limited supervision, extract, and organise knowledge out of unstructured data, analyse discourse for improved contextual understanding, and many more.
KENOME Insights Platform (KIP) is one such platform that is designed in such a way that it digs out actionable insights from dark data. The platform uses machine learning, deep learning, and natural language understanding to build enterprise knowledge graphs out of unstructured data. Also, the platform provides inference capabilities over such graphs for improved decision making.
Here are some of the products built on top of the KIP framework:
- KENOME SmartOrder helps enterprises create sales orders from heterogeneous sources, such as chat messages, emails, SMS, etc.
- KENOME SmartAudit helps companies generate revenue audit leads from heterogeneous data sources. This is a prescriptive system which not only generates revenue audit leads but also presents supporting evidence for them.
- KENOME SmartRec uses Knowledge Graphs to make personalised recommendations while taking external trends and other Social Sensing signals into account. SmartRec also exploits semantic relationships among products.
- KENOME SmartSense derives Social Sensing signals by continuously scanning a variety of heterogeneous sources where customers express their opinions.
- TRAKCrypto is built on top of KENOME SmartSense to track cryptocurrency prices using heterogeneous unstructured data and deep learning. This is the first AI-based crypto price estimation tool of its kind.
Staying Ahead Of The Curve
Talking about competition, the use of Knowledge Graphs to make sense of enterprise dark data is in its infancy. Lattice Data was a startup in this area which was acquired by Apple. Big players such as Google and Microsoft have built knowledge graphs for the general web and are using them for web search. And currently, Amazon is developing a Product Knowledge Graph.
KENOME is one of the first startups to use knowledge graphs, bringing years of world-class research to address challenging but unexplored problems around unstructured data. Also, the company is in continuous learning and improving phase in terms of both products and features based on the feedback from customers.
“Success of our products with the first set of customers has given us the confidence to take these offerings to other customers to get further validation and feedback. We are actively working on that and are very excited about the journey we are on,” Partha concluded.