Office work usually doesn't break because people don't care. It breaks because the client asked in WhatsApp, the task lived somewhere else, and no one could see both at the same time.
This week we shipped a clearer path: setup can bring in older office chats, each task keeps the sender and message context, and approval steps are now written for humans instead of internal shorthand.
The goal is simple: fewer lost requests, less copy-paste, and one place to see what the client asked for, who owns it, and what still needs action. Early pilot, still tightening it. Link in replies.
The first WhatsApp task board just shipped in Agente Desktop. Office managers see WhatsApp messages from team accounts turn into an organized task board โ in one place, no copy-paste. The computer files. The human approves.
Every WhatsApp conversation now has its own listen scope per account. Switch between them in the communication dock. Messages route to the right board. Nothing gets lost.
After onboarding, users land on the live ticket board โ not a blank screen. This sounds obvious. In office software, it is not. First impressions determine whether the team keeps using it or walks away.
The entire product got a unified visual identity. Shared design tokens now connect the login screen, sidebar, onboarding, and landing page. One product that looks like one product โ from first visit to daily use.
I kept seeing the same problem with companies trying to adopt AI:
They usually do not need another random tool. They need to know where AI would actually help first.
So I built a quick AI Maturity Audit for Crux.
It scores a company across 6 areas: strategy, adoption, governance, workflow integration, augmentation, and feedback loops. Then it gives you the strongest area, the biggest gap, and a few next actions.
Hereโs a sample report I ran:
https://t.co/ejE9Km4dpb
And if you want to run it for your own team, the audit is here:
https://t.co/6b3whaG8aL
The goal is simple: move from โwe should be using AI moreโ to โthis is the workflow we should improve first.โ
Really excited to introduce IterVox!
Multi agent and multi machine orchestrator. Fully open source.
It started few weeks ago as implementation of @OpenAI Symphony spec for Claude code, but since then evolved quite significantly into multi agent orchestrator with tons of features.
Both @claudeai and @OpenAI Codex are supported.
@linear and @github issues integration
Check it out and lmk what do you think
https://t.co/z5VSytw8CV
Been using Paperclip from @dotta for a week and it genuinely feels like the future.
https://t.co/dOS4yVs2C0
Not because it gives you another AI agent.
Because it gives you a company of agents.
The CEO is an agent.
The CTO is an agent.
The departments are agents.
And you sit above them as the board: setting goals, approving direction, and watching execution run 24/7.
Thatโs the shift:
from asking an assistant to do work
to governing an organization that does.
And the scary part is that this already feels more limited by model caps than by the orchestration itself.
@OpenAI released the Symphony spec for autonomous AI agents.
I built the Go implementation โ for @AnthropicAI Claude code
Watch Claude Code picks up Linear tickets, write code, open PRs, and
move your Kanban board while you sleep.
https://t.co/Y0nROmGtbs
Install with
brew install vnovick/tap/symphony-go