We’re excited to announce Formant’s new integration with generative AI, Theopolis.
ChatGPT is the leading large-language model (LLM). Available today, Theopolis provides both the standardized structured Formant workflows enabling users to manage robotic fleets at scale, as well as the new exploratory workflows that allow users to ask exploratory questions of their robot data to receive immediate insights.
Our goal as a platform is to unlock the potential shared by humans and robots with more efficient and flexible workflows. By allowing GPT-4 to apply its own creativity to a given data set rather than having a human do the same, we’re saving robotics companies time spent configuring operator dashboards and making their data more valuable than ever before.
Anyone can be a data scientist
It’s easy to see when a robot is experiencing an issue; determining why the issue has occurred is a much more difficult and time-consuming task. By harnessing AI’s abilities to evaluate data, it becomes much easier to find root cause of issues, gaps in performance, and opportunities to optimize.
Using basic SQL queries with the Snowflake data schema provided in Formant, Theopolis can provide information about current data from a fleet, suggest SQL queries and output well-lined JSON with valid JSON escape codes, and suggest appropriate chart types and chart-specific details based on the query output.
For example, an operator can ask Theopolis, “Are there any questions a data scientist may ask about {insert table name]”, and receive a list of generated prompts to help them build further questions, such as “Please show me [ insert generated prompt] as a pie chart”. Theopolis is then able to generate a chart for the question using analytics pulled from the fleet, saving operators the time spent to generate prompts themselves, and providing a creative advantage.
Taking a cautious approach to AI
Controlling a robot with LLMs isn’t safe yet, but analyzing data that is understood by them is. We take AI seriously at Formant, and believe that any step forward with the technology should only be done in total confidence.
To ensure this value, Theopolis focuses on data presentation rather than any physical actions, and offers only white-listed commands that must be confirmed by a user before executing. This prevents a situation where a free range of robot actions and behaviors leads to unintended consequences.
Looking forward
Formant is uniquely positioned as the interface that links humans to robotic fleets. We’re excited to provide a feature that will further help robotics companies leverage their data and the value of using AI to optimize the efficacy and performance of their robot operations.
Our team looks forward to helping robotics companies accelerate their success with the safe use of GPT-4, and will continue to take steps to ensure confidence in any usage of AI within the Formant platform.
Visit our website to learn more about Theopolis, or request a demo to meet with our team.
