We are committed to building AI systems that provide measurable value. Our solutions build on the four foundations of effective AI: original research, collaboration, engineering, and evaluation.
Four Foundations of Effective AI
ORIGINAL RESEARCH: leading the way
Our team has years of first-hand research in automatic question-answering, natural language processing (NLP), image and speech processing, information retrieval, and statistical machine learning. We know state-of-the-art methods, which we combine and improve in novel ways.
COLLABORATION: identifying opportunities
Our clients and partners are true collaborators, contributing their deep knowledge of the context, in which an AI system can be deployed. This level of partnership is how we ensure our solutions address real issues and rely on the right data, analytics, and technologies. We partner with you; we don’t just sell you a product.
ENGINEERING: putting it all together
Our team of AI research leaders and skilled implementers bring it all together, creatively and efficiently combining advanced AI components to build applications that learn from your data and respond to the complexity and unpredictability of today’s world. We customize a solution to your situation, not the other way around. So far we have developed adaptive AI applications in 3 areas: Cognitive Insights, Cognitive Engagement, and Cognitive Information Agents. See below for more about our emerging product areas.
EVALUATION: measuring impact
We are committed to rigorous assessment of our systems’ performance using both intrinsic and extrinsic evaluation. In intrinsic evaluation we rely on various proxy quality measures (e.g., accuracy of speech recognition, or task completion rate of a conversational agent). We also implement extrinsic evaluations to assess our system’s impact on our client’s progress towards their goals (e.g., revenue, store visits, rehospitalization rates). To this end, we rely on randomized controlled experiments on data in the field.
Emerging Product Areas: Multi-Strategy, adaptive AI “in the field”
We have developed multi-strategy, adaptive AI systems in three areas: Cognitive Insights, Cognitive Engagement and Cognitive Information Agents. These emerging AI applications are already humanizing data, augmenting technology, and empowering people.
Cognitive Insights distils large and unstructured data stores to present valuable information that you didn’t even know to look for. In addition to supporting client-led inquiries, these systems can automatically identify patterns, anomalies, or cross-compliance concerns in human interactions through organizing data around dynamic clusters. Applications of these systems are in threat detection for cyber security, risk detection for finance and insurance, or health and security alerts.
Cognitive Information Agents
Cognitive Information Agents speak to the piece of the AI puzzle that is “acting like a human.” Chatbots, voice interfaces, conversational interfaces — all of these applications recognize and interact with people as if they were having a conversation. Combined with the right analytics, these interactive and contextually-aware agents learn user preferences and provide evidence-based answers — without human intervention. Cognitive Information Agents have endless use cases: help desk support, enterprise software agents, point-of-sale systems and so much more.
Our Cognitive Engagement system continuously monitors transactions, communications between buyers and sellers, and social media activity of individual buyers across markets. Real-time analysis of cumulative market participation and communication identifies contextually relevant recommendations, effective offers, and preferred communication channels that companies can use to generate impactful, individual-level content for customers. Applications of these AI systems can increase efficiency when nurturing long-term customer relationships in retail trade and service industries.
What’s your challenge?
Bring your challenge to the table and we’ll apply our adaptive AI toolkit to craft a tailored solution. The technology is flexible and scalable, so no problem is too big or too small for our product developers and data scientists.