AeroOps - AI Operations Manager

Inspiration

AeroOps — AI Operations Manager Track 3 — AI-Native Enterprise (Open) The Problem In many companies, when an operational problem occurs, employees have to manually monitor different systems, understand what went wrong, decide what needs to be done, coordinate with the right people, and finally check whether the problem was actually solved. Existing dashboards can show that something is wrong, but they don't always help answer the more important question: "What should we do next?" We wanted to build an AI system that could help with this entire process.

What it does

Our Solution AeroOps is an AI Operations Manager that takes a user's operational goal and works through the problem step by step. Instead of asking an employee to decide every action, the employee can simply describe what they want in natural language. For example: "Monitor Project Alpha and handle anything affecting delivery." AeroOps then follows an agentic workflow: Understand → Plan → Gather Information → Select Tools → Execute → Verify → Report The agent identifies what needs to be done, selects the appropriate tools, performs the required actions and checks whether the expected result has been achieved. For actions that may have a significant impact, AeroOps includes a human approval step before proceeding.

How we built it

How Our Prototype Works We created an interactive web prototype called the AeroOps Command Center. The prototype includes: Command Center — where the user provides an operational objective. AI Agent — demonstrates the agent's step-by-step execution workflow. Monitoring — displays operational signals, risks and anomalies. Tasks — shows tasks created as part of the workflow. Tools & Integrations — demonstrates the tools that the agent can use. Human Approval — allows a person to approve or reject an important action. Reports — provides the final execution result. Execution Trace — shows the stages followed by the agent.

Challenges we ran into

In our demonstration, AeroOps receives a project delivery objective, identifies a delivery risk, investigates the issue, determines the required action and verifies the outcome. When a high-impact action requires approval, the workflow pauses for human confirmation

Accomplishments that we're proud of

What Makes AeroOps Different Our goal isn't to build another dashboard that simply gives employees more information. We want AeroOps to act as an operations partner that can take a goal, figure out the steps required and help complete the work.

What we learned

At the same time, we don't want an enterprise AI agent to operate without control. That's why our design includes human approval for important actions and an execution trace so users can understand what happened during the process.

What's next for AeroOps - AI Operations Manager

Future Scope If we continue developing AeroOps, we want to connect it with real enterprise systems such as project management platforms, ticketing systems, communication tools and operational monitoring systems. We also want AeroOps to become more proactive by detecting potential problems before they become major issues, while maintaining human approval for important decisions.

Team Swarna Rekha — Team Lead AI Agent & Product Development I worked on the AeroOps concept, agent workflow and interactive prototype. I focused on how the agent should understand a user's goal, determine the required steps, use tools and verify the final result. Nitin George — Team Member Research, Operations & Validation Nitin worked on researching the enterprise operations problem, defining operational scenarios and workflows, testing and validating the prototype experience, and supporting the documentation and demo preparation.

Built With

• css
• githubpages
• html
• javascript
• vanilla
• vscode