Geeknewz exclusive comparison built from Ando's launch coverage, Block's Buzz announcement, and how Slack and Microsoft Teams currently treat AI agents as add-ons rather than coworkers.
Sara Du spent 2025 helping companies stand up MCP servers so their AI agents could talk to tools. The question she kept hearing was simpler and more awkward: how do we use those agents inside Slack without turning a human into the messenger? Her answer, Ando, came out of stealth on September 24 with $20 million from Accel, Index Ventures, and Emergence, and a blunt thesis. Slack and Teams were built for people, while agents are becoming participants. Treating them like plug-in bots creates what she calls "meat proxies," humans who copy an agent's output into the channel so everyone else can see it.
That framing is not unique to Ando. Block shipped Buzz in July as an open-source, Nostr-based workspace where humans and agents share channels with cryptographic identities. Slack has been promoting its own AI agent. Microsoft keeps folding Copilot deeper into Teams. The race is less about chat UI and more about who gets to define how agents sit in the room.
What Ando actually ships
Ando is pitching a full Slack replacement for teams that already run agents. Agents get their own identities and inboxes. They can browse channels, join conversations without being tagged, sit in on transcribed live calls, and message a human when they decide judgment is needed. Du told TechCrunch the product felt "jankier" than Slack early on, which is an honest admission, and that retention improved once customers saw an agent notice two related threads and pull people together without being asked.
Customers today span software, real estate, and finance across about 15 countries, mostly small teams. The fresh capital is for hiring and, in Du's words, burning through more tokens. That last detail matters, because an agent-native chat product lives or dies on context windows and API bills as much as on message search.
Side-by-side: four bets on the same problem
| Product | Agent role | Hosting / lock-in | Stage | Best fit (Geeknewz view) |
|---|---|---|---|---|
| Ando | First-class coworker with inbox, channel browse, proactive DMs | Proprietary SaaS; $20M seed/pre-seed | Early customers, small teams | Startups already drowning in agent handoffs |
| Buzz (Block) | First-class coworker with portable crypto identity | Open source (Apache-2.0), self-host or Block-hosted; model-agnostic | Public since July 2026 | Teams that want agent identity they own |
| Slack AI agent | Bolted-on assistant inside existing Slack | Slack / Salesforce cloud | Mature platform, evolving agent | Companies that will not rip out Slack |
| Microsoft Teams + Copilot | Assistant woven through M365 | Microsoft cloud, Office suite gravity | Enterprise-default | IT shops standardized on Microsoft |
The useful split is architectural. Ando and Buzz both treat agents as members. Slack and Teams treat agents as features you enable on a human-first product. Buzz's Nostr identity pitch is the most radical: the agent keypair belongs to the agent, not the vendor. Ando's pitch is product speed and UX for the same idea without asking you to run a protocol. Incumbents win on integrations, compliance, mobile apps, and the sheer cost of migration.
Geeknewz verdict
Should you rip out Slack this week? Almost certainly not, unless your team is small, agent-heavy, and already frustrated by copy-paste "meat proxy" workflows. Watch Ando if you want a polished commercial bet on agent-native chat and can tolerate early-product rough edges. Watch Buzz if you care about open infrastructure and portable agent identity. Stay on Slack or Teams if your real problem is still search, permissions, and SSO, not agent participation. The interesting signal is that two well-funded challengers now agree Slack's bot model is the wrong abstraction. That does not mean Slack loses. It means the next five years of workplace chat will be argued in agent terms, not emoji reactions.