The decade of the 2010s has seen artificial intelligence (AI) flow out of universities and research labs into mainstream commercial deployment, specially involving image, speech and natural language processing in large consumer Internet platforms. We believe that it will be the key driver for innovation in enterprise business operations going ahead. Recent advancements in AI techniques have evolved its linear statistical learning abilities (“machine-learning”) into scalable & complex feed forward multi-layer learning systems that can accurately model non-linear real world scenarios (“deep-learning”). There has also been significant maturing in AI workflow management allowing the decoupling of large scale model training from model selection & serving, making it feasible to offer AI-driven services in a convenient SaaS-based deployment model. The unit economics of cloud-based GPU utility compute has made large scale AI training economically feasible. Applications of AI in process automation, autonomous systems and business operations in the enterprise are a key focus area for The Hive.


Conversational agents are computational systems and services built to converse with a human with a coherent dialogue structure across text, speech, graphics, gestures and other modes of communication. The explosive growth of mobile commerce driven by cloud-hosted consumer services has led to the development and deployment of large-scale conversational agent platforms like Apple Siri, Facebook Messenger, Google Assistant, Microsoft Cortana, Kik and Amazon Alexa for enhanced conversational social collaboration, discovery and user-experience. These platforms use natural language processing (NLP) and natural language understanding (NLU) techniques to deliver coherency in conversations. The efficacy and accuracy of conversational systems have been greatly enhanced over the recent years due to advancements in computational semantics and deep linguistic processing. The Hive’s Synapse Application Stack drives rich conversations with multiple turns for domain-specialized conversational agents by offering a combination of rich semantic modeling and deep learning. While having roots in consumer internet, conversational agents are disrupting the interfaces between people and processes in enterprise business operations. Business applications of conversational agents in areas like customer service & support, business collaboration and B2B commerce are one of the key focus areas for The Hive.


The late 2000s saw massive consolidation of compute capacity through the rapid growth in utility cloud computing for consumer and enterprise applications. This trend has been counter-balanced by the large-scale deployment earlier this decade of intelligent compute devices on a person (phones, wearables), and Internet of Things in homes, workplaces and industrial sites. This has led modern consumer and IoT applications to go beyond a cloud-only deployment model to utilizing compute resources available at the Edge of its interface with the clients. The key advantages of such edge intelligence include better economics of data transmission (specially in high-volume, high-velocity scenarios), enhanced data security & privacy features, real-time responsiveness, ability to capture context from ambient physical environments (called “ambient intelligence”), and higher levels of business continuity (due to distributed service deployment). Applications of edge intelligence (in different contexts also known as Ambient Intelligence or Fog Computing) in industrial operations, smart-infrastructure and datacenter operations will continue to be one of key focus areas for The Hive. Past co-creations within The Hive have explored applications of Edge Intelligence in Industrial IoT (FogHorn), IoT security (Sensify Security) and consumer IoT (Snips).


The post-2008 world has seen a severe regulatory clampdown and increase in overhead of compliance on businesses. Blockchains present a new information system that manifests multilateral trust & transparency. These are decentralized systems of record or distributed ledgers, which can efficiently enforce multilateral consensus between all counter-parties on every transaction recorded. Core systems in today’s enterprises use relational databases with unilateral consensus across a given institution, and deploy reconciliation, settlement and audit services to build consensus across institutions. Blockchains break down these large batch services by systemically incorporating them in a digital protocol based consensus that includes all counterparties and stakeholders of the transaction. Blockchain protocols also build redundancy to tolerate malicious counterparties within pre-designed threshold limits. The Hive is exploring business applications built on blockchain networks addressing usecases in financial services, insurance, commerce, supply-chain management and industrial operations.


In the recent years, we have seen computer vision technologies commercially deployed in consumer applications due to commoditization of cameras in mobile devices, as well as applications in autonomous vehicles, medical imaging, etc. Applications of deep learning techniques like convolutional neural networks (CNNs) in the recent years have brought accuracy levels of object recognition in images to higher than that of human cognition. Since then, the research community has been making great strides in scene recognition through automation of image feature extractions and labeling. Computer image rendering for mobile image, video and AR/VR visual perceptions has been scaled up by recently launched cloud rending services that utilize utility GPU processing farms. All these developments have matured computer vision for production enterprise applications. The Hive focuses on business applications of computer vision in eCommerce, autonomous navigation, logistics & transportation, enterprise automation and industrial services & process management.