At icCube, we are embarking on a fundamental shift in how our users interact with our product. We are moving away from manual, time-intensive workflows toward a highly interactive experience where AI acts as an expert-partner in the development and analysis process.
So, how are we actually going to do this? Our AI roadmap focuses on four key areas designed to cut out the manual dashboard, MDX and schema work so you can focus on driving real value:
- Chat with Your Dashboards: Your end-users can simply ask questions in plain language, and the AI will serve up the right insights and visualizations instantly.
- AI-Powered Development: Think of this as a coding co-pilot inside icCube. We’re integrating AI to help you with bug fixes, MDX generation, and schema validation. You get real-time feedback so you can directly validate the output of the AI.
- Automated MDX Generation: By automating MDX generation through natural language, we’re lowering the technical barrier, letting you build faster without needing to be a query expert.
- Intelligent AI Assistants: We’re giving our agents access to our own technical documentation. This means the agent knows how icCube works and thus can better solve the problems you give it.
Architecture
We’re moving past simple "chat" interfaces toward a true agentic architecture. The goal here is to give the AI the right tools so the answers it gives are valid and correct.
- MCP Integration: We’re using Model Context Protocol (MCP) servers to connect our tools (like our dashboard editor and schema builder) directly to the LLM. This allows the AI to make changes in-place while you’re working in the dashboard editor or the model builder.
- Separate AI instance: The agents will live in their own containerized environment. From there, it has access to icCube via the tools we provide it. These tools include:
- MDX query validation
- MDX query execution
- Schema discovery
- Builder: schema validation
- Dashboard editor: dashboard validation
- Human-in-the-Loop: For the editor, we’re building this with a "human-in-the-loop" approach. The AI handles the heavy lifting, but your team validates the quality of the result. We want production-ready schemas and dashboards, not just prototypes.
What’s on the Horizon
We’re in an aggressive innovation phase, but our goal remains simple: agents that give you valuable data insights and speed up how you work with icCube. This allows you to combine your subject expertise with icCube’s AI-agent to help you build and iterate faster than ever before.
We’ve got plenty more in the works, and we’ll be sharing more as our roadmap matures.
If you want to be part of these developments and have early access, please let us know. We’re eager to test the agent for your specific use cases!