ServiceNow Research
Empowering responsible AI for efficient workflows
About
ServiceNow Research, established in 2003 and based in Santa Clara, California, is an integral component of the ServiceNow organization, dedicated to pushing the boundaries in artificial intelligence (AI) research. Its mission to enhance AI-powered experiences is evident in its focus on both fundamental and applied research, particularly in the context of the Now Platform. A notable aspect of ServiceNow Research is its emphasis on future-proofing AI experiences by making workflows smarter and AI workloads more efficient. This forward-thinking approach extends to the responsible development of AI systems, prioritizing values such as fairness, ethics, accountability, and transparency. A considerable portion of ServiceNow Research's work revolves around generative AI and large language models (LLMs), showcasing its commitment to advancing these cutting-edge technologies. They actively participate in a variety of research programs, such as the Conversational Assistant program, which aims to create reliable and trustworthy AI assistants. Moreover, their work in Human Decision Support explores AI's capabilities in fields such as causality, forecasting, and planning, thereby aiding human decision-making processes. Additionally, their Multimodal Foundation Models program seeks to integrate multiple data types—ranging from text and images to tabular data and graphs—into foundational models with expansive capabilities. ServiceNow Research has made significant contributions to AI literature and practices. For instance, they have published work at the NeurIPS Datasets and Benchmarks Track, which includes initiatives like RepliQA, focusing on benchmarking LLMs using unseen reference content. Their research on multimodal foundation world models for generalist embodied agents and the responsible development of general-purpose AI models, particularly LLMs, highlights their dedication to advancing the field. These efforts are complemented by their initiatives like BigCode and involvement in open-source projects, making vital tools and datasets available to the broader AI community. Further strengthening its AI expertise, ServiceNow acquired Element AI, enhancing its capabilities and research prowess. The company also collaborates with industry leaders like NVIDIA to harness high-demand GPU computing power essential for LLM training. ServiceNow’s capability in generative AI is exemplified through the Now Assist platform, which integrates generative capabilities to streamline and augment various workflows. By leveraging general-purpose LLMs from services such as Microsoft Azure OpenAI and developing their own domain-specific models, ServiceNow ensures accuracy, cost-effectiveness, and superior user experiences. Enhanced productivity features such as case summarization and text-to-code conversion, along with the development of Now LLM for enterprise scenarios, further underline their commitment to specialized and effective AI solutions.
