Agentspace

A centralized platform that enables users to communicate with multiple AI agents through a single interface while monitoring their activity, performance, success rates, and operational costs in real time.

problem

As I expanded my ecosystem of AI agents, I found that managing them through Slack created more friction than visibility. Conversations were spread across channels, operational data was difficult to access, and there was no clear way to understand what each agent was doing, how much it was costing to run, or whether it was successfully completing tasks. While Slack provided a communication layer, it lacked the observability needed to effectively manage and scale a growing multi-agent system.

solution

I designed and developed a centralized Agent Operations Platform that combines a unified chat experience with real-time operational monitoring. The platform allows me to communicate with all agents from a single interface while providing visibility into execution status, success rates, token consumption, and overall operating costs. By bringing communication and observability together, I created a single source of truth for managing agent performance.

The project began as a personal frustration. As I continued building more agents, I realised that I spent more time switching between tools and investigating behaviour than actually improving the agents themselves. Critical information was fragmented across conversations, logs, and external monitoring tools, making it difficult to understand the health of the overall system.

I approached the challenge by thinking beyond a chat interface and instead designing an operational layer for AI agents. Rather than simply providing a way to interact with agents, I wanted to create a platform that exposed the metrics and telemetry needed to understand how they were performing in production. This led to the creation of a dashboard that aggregates agent activity, execution data, costs, and performance metrics into a single view.

From a technical perspective, the platform collects telemetry from each agent and surfaces it through a real-time monitoring dashboard. This enables visibility into execution outcomes, operational costs, and system performance, while the unified chat interface provides a streamlined way to communicate with agents without leaving the platform.

The result is a scalable control centre for managing AI agents. What started as a solution to my own workflow challenges evolved into an operational platform that provides transparency, control, and insight across an entire multi-agent ecosystem.

year

2026

timeframe

16 days

tools

Figma, Claude code, Cursor