An operating system for how work gets done
The problem
A growing IT consulting firm was running a 50+ client portfolio out of Asana for tasks, Slack for conversations, and Notion for knowledge. Each tool worked fine in isolation. Together they didn't. Work came in through five channels, lived across three systems, and required someone on the team to manually triage and route everything. Things got lost. Context got duplicated. The team spent more time switching tools than doing the work itself, and the operating model clearly wasn't built to scale with the client base they'd already won.
The engagement
Build a unified operating layer on top of the existing tools rather than ripping them out. One intake point for every channel and team. Intelligent triage that routes by team and type. And AI agents pulling the strings underneath, resolving routine questions before they ever became tickets.
What we did
We built a shared operating platform that sits across Asana, Slack, and Notion and treats them as one connected layer rather than three disconnected ones. Shared data, shared auth, shared identity. The team kept the tools they already knew. We added the connective tissue.
On top of that layer, an AI intake agent lives wherever the team already works, primarily in Slack. It's a single chat surface that classifies incoming work as it lands, routes by team and type, and creates structured tickets in Asana automatically, with the right metadata and the right owner attached from the start.
Underneath the chat, domain-specific sub-agents run invisibly: search, ads, data, knowledge. A routine question goes to the right sub-agent and gets answered in seconds, often with the relevant Notion doc cited inline. Real work becomes a ticket in Asana with metadata, owner, and context already attached. The team stopped routing work and started shipping it.
The impact
Six tools consolidated into one operating layer with shared auth and shared data. Manual ticket routing eliminated entirely. ~60% of routine intake now resolves through the sub-agents before a human ever sees it, so the queue is shorter and the work that lands on a consultant is higher-leverage. With routine work absorbed, each consultant now supports roughly 30% more client accounts without added headcount, which is the economic outcome the firm was actually after. And the platform gives them the substrate to layer future AI capabilities on top, without rebuilding the foundation each time.