OnticWorks.io uses IBM Bob to connect AI reasoning with the contexts in which people work, operate and learn
OnticWorks.io, the AI and robotics subsidiary of Avid Solutions International, develops agentic systems designed to bridge digital intelligence and physical autonomy. The company was founded around a simple idea: people experience work through place, context and relationships, while most autonomous systems reason through abstractions of those experiences. To help close that gap, OnticWorks.io developed Elmer, an agentic operating system that combines voice, vision, gesture recognition and operational data to help users understand situations, evaluate options and take action within context.
Traditional systems often translate real-world conditions into dashboards, reports and data models before applying intelligence to them. As a result, professionals must interpret recommendations and reconnect them to the environment in which they need to act. Important signals can remain buried in email threads, sensor alerts or simulation outputs, slowing decisions and increasing the effort required to move from insight to action.
Those limitations became particularly visible in three environments where OnticWorks.io was applying AI and automation: administrative operations, agriculture, and robotics education. Although the environments differed, each required the same capability: understanding what was happening, where it was happening and what action might be needed next.
Administrative and operational professionals relied on manual email triage, scattered notes and memory-dependent follow-ups, with limited visibility across calendars, opportunities and day-to-day priorities. Important requests, customer communications and emerging opportunities could remain hidden within growing volumes of information, forcing teams into reactive rather than proactive decision-making.
Agricultural operators faced a more physical version of the same disconnect. Field monitoring required walking crop rows and recording observations manually. Irrigation schedules relied on historical patterns without real-time integration of sensor information, and crop stress could go undetected until its effects became visible.
In education, RoboLabWorks, a robotics learning environment developed by OnticWorks.io, was working to bridge one of robotics training’s most persistent divides: the gap between succeeding in simulation and deploying safely to physical hardware. Students could successfully complete robotics exercises in virtual environments, yet still face significant challenges when moving to physical systems, where unpredictable conditions, sensor noise and safety constraints introduced complexities that did not exist in controlled simulations.
Across offices, fields and classrooms, people still had to bridge the gap between what they were experiencing and what the system understood, leaving intelligence one step removed from reality.
To move intelligence closer to the environments in which decisions occur, OnticWorks.io integrated the spatial understanding capabilities of IBM Bob™ into Elmer’s core reasoning loop.
Rather than processing information as isolated inputs, Elmer was designed to understand relationships between people, places, assets and events as they unfold in real time. As Elmer processed information from cameras, sensors, enterprise applications and user interactions, Bob helped provide the spatial context needed to understand how people, places, assets and events related to one another. Combined with IBM watsonx.ai® for multimodal reasoning, NVIDIA-powered inference capabilities, and deployed on Red Hat® OpenShift®, Elmer could move beyond analyzing data about the world to reasoning within it. Prior to the integration, Elmer primarily relied on dashboards and separate workflows across digital and physical environments. With spatial context incorporated directly into the reasoning process, the platform could continuously adapt its recommendations as conditions changed.
The resulting framework combined voice, vision, gesture recognition and operational telemetry through a multimodal interface. Information from webcams, environmental sensors, drone systems and enterprise applications could be considered together before Elmer proposed a recommendation or action. Every recommendation followed the same draft-and-approve workflow, ensuring that users remained in control of consequential decisions while continuously teaching the system through approvals, modifications and overrides.
For administrative professionals, Elmer connected email, calendars, opportunities and environmental context to help prioritize actions according to their relevance and urgency. In agricultural operations, the platform combined drone positions, soil conditions, weather data and equipment telemetry to reason about operational constraints before proposing irrigation, routing or treatment decisions. For RoboLabWorks students, the same framework helped connect simulation and physical deployment, allowing learners to understand failures and outcomes through real-world context rather than abstract error logs.
Rather than building separate products, OnticWorks.io extended the same agentic framework across productivity, agriculture, and robotics education, supporting founders, farmers and students through a shared understanding of place and context.
Moving from analyzing information to reasoning within context delivered different outcomes across each environment, but each reflected the same underlying shift toward bringing intelligence closer to where work happens.
For administrative professionals, Elmer helped surface relevant context more quickly, reducing the effort required to prioritize actions and make informed decisions. Managers reported saving between eight and twelve hours each week on email triage, briefings and calendar coordination, while gaining faster access to the information most relevant to their priorities.
In agricultural operations, spatially grounded decision-making contributed to measurable field outcomes. Irrigation-managed deployments achieved a 58% reduction in water use while improving crop yields by 23%. Labor devoted to routine monitoring and field decision-making decreased by an estimated 40%, reducing the need for manual scouting and allowing personnel to focus on higher-value operational activities.
For RoboLabWorks, realistic simulation environments and spatially grounded learning accelerated development and deployment. The platform’s launch timeline advanced from 2027 to 2026, representing a 12-month acceleration. In addition, 24 of 27 participating students successfully deployed autonomous agents to physical hardware, demonstrating stronger transfer from simulation-based learning to real-world operation.
While the operational improvements were substantial, OnticWorks.io viewed the project as validation of a broader approach to agentic systems. By moving Elmer from processing information about the world to reasoning within it, OnticWorks.io extended a single agentic framework across offices, fields and laboratories through a shared understanding of place, context and human oversight.
Avid Solutions International helps organizations apply technology to complex operational challenges through engineering, automation and AI-driven solutions. Through its AI and robotics subsidiary, OnticWorks.io, the company is extending that expertise into agentic systems that connect digital intelligence with physical environments and help enterprises transform operations through intelligent automation.
Legal
© Copyright IBM Corporation September, 2026.
IBM, the IBM logo, and IBM Bob™, IBM watsonx.ai®, Red Hat® OpenShift®. are trademarks of IBM Corp., registered in many jurisdictions worldwide.
Examples presented as illustrative only. Actual results will vary based on client configurations and conditions and, therefore, generally expected results cannot be provided.