The Center for Data Innovation spoke with Raghu Bala, founder and CEO of Synergetics AI, a California-based startup that builds AI-enabled virtual assistants that can automate professional tasks for organizations. Bala discussed how Synergetics’ AI-powered assistants can independently collaborate with each other and how future quantum computing advancements will allow the company to improve its AI assistants.
Martin Makaryan: What does Synergetics AI offer?
Raghu Bala: Synergetics provides virtual assistants that use AI to streamline and automate various administrative and professional tasks in organizations. These virtual assistants can perform many of the tasks that human employees can. We built a platform that allows our clients to fine-tune large language models with their own data, customizing these AI agents to specific tasks. Agent-to-agent communication is a key feature of what we offer. It allows virtual assistants from different enterprises to identify, discover, and communicate with each other across boundaries. In other words, an AI assistant in one company is able to collaborate with a virtual assistant from another organization, similar to how human employees would. This is crucial for decentralized autonomous agents that operate beyond a single enterprise’s confines.
Makaryan: What sets Synergetics apart from other startups?
Bala: Unlike many AI companies that focus on single, specialized assistants solving specific problems, we emphasize workflow integration across multiple roles within an organization. Our ability to support complex workflows and enable decentralized autonomous agent communication distinguishes us from many competitors. This capability positions us well in the market as more companies look to integrate sophisticated AI solutions that go beyond simple task automation.
Makaryan: What are examples of client use cases for AI agents?
Bala: Our clients primarily use our AI agents within their enterprises to automate processes in healthcare and financial services. For instance, in healthcare, we have developed agents that provide counseling for patients struggling with substance addiction and manage digital medical records. These agents streamline data entry and improve productivity by reducing the time clinicians spend on administrative tasks. In financial services, we have co-developed agents with a bank for wealth management and customer service.
Makaryan: What models power Synergetics’ AI assistants?
Bala: Synergetics uses various models to power our AI technology, including out-of-the-box, pre-trained models from large developers like OpenAI. The platform allows users to upload and fine-tune their own models using multiple data types, including images, video, audio, text, and PDFs. Synergetics employs a layered approach to training, starting with broad industry data and narrowing down to specific customer data. This approach helps minimize the risk of hallucination in outputs. During the inference phase, we use techniques like retrieval augmented generation to further refine the AI assistant’s responses based on specific customer requests. Synergetics’ platform is user-friendly and offers tools that make it easy for users to train and deploy AI models without coding knowledge. Once these customized models are ready, customers can deploy them in the form of virtual AI assistants, which can complete various tasks. The objective is to automate many time-consuming functions that humans currently complete and ensure communication channels between these virtual assistants so that they can be as efficient as possible.
Makaryan: What future advancements do you foresee impacting Synergetics?
Bala: I think something that is not on the radar of many people is quantum computing, which I believe will significantly impact our field by accelerating AI processing capabilities. As quantum technology matures, it will enable faster computations and more efficient data handling, enhancing the performance of our AI agents. We anticipate that by the late 2020s, quantum computing will become more prevalent and economically feasible, allowing us to leverage the increase in computing power to improve our AI assistants.
