Fourteen Places Your Agent Can Answer Besides the Chat Window
A full accounting of the messaging platforms a Piskes agent can be reached on natively, without stitching together separate point tools.
Before you evaluate whether an AI agent fits into how your team actually communicates, the real first question is usually simpler: where does it live? Most teams don't want to open a new tab every time they want to ask their agent something — they want it where they already are. Piskes ships that as a native list of channels, not a promise to "integrate with your stack eventually."
The list
The Channels screen inside ai-agent connects your agent to fourteen separate messaging platforms, each backed by its own dedicated worker: Slack, Discord, Telegram, WhatsApp, Microsoft Teams, DingTalk, WeCom, Feishu, Matrix, Mattermost, ntfy, Twilio (SMS), Line, and Google Chat.
That's a genuinely unusual breadth for a single product to ship natively. Most teams evaluating "can my agent live where my team talks" end up stitching together three or four separate point tools to cover this ground — a Slack bot from one vendor, a Telegram integration from another, custom webhook glue for anything WhatsApp- or Teams-shaped. Piskes's approach is to build the adapters directly, so the coverage doesn't depend on what a third party happened to prioritize.
What "connected" actually means, per channel
Each channel adapter has to be understood in two directions, and they're not automatically the same thing: can your agent receive a message on that platform, and can it send one back. A channel being listed as available doesn't automatically mean both directions are equally mature for every platform — check the specific channel's status on the Channels screen rather than assuming uniform coverage across all fourteen, since messaging platforms differ in what their own APIs make easy or hard.
An honest note on the count
The platform's own internal accounting describes more adapters as deployed than have been independently confirmed reachable — the code that powers this screen notes that a number of adapters are deployed while a slightly smaller number have actually served a live health check. The fourteen named above are the ones that can be confirmed directly from the deployed worker list, and that's the number worth relying on until a fifteenth is independently verified and named publicly. If your team is evaluating a platform not on this list, it's worth asking directly rather than assuming — this is a fast-moving surface.
Why this matters more than it looks like it should
The messaging-platform-deployment conversation in 2026 keeps circling the same theme: comparisons of AI bots across Telegram, Slack, and Discord, and pushes toward "cross-platform messaging" as a first evaluation criterion for any AI tool a team is considering. The unstated assumption behind most of that coverage is that you'll need multiple tools to actually get there. Piskes's bet is that if your agent already lives inside a full business platform — CRM, Drive, tasks, and the rest — it should also be reachable wherever your team is already talking, without you having to build that bridge yourself.
If you're getting started, the practical move is simple: go to the Channels screen, find the platform your team actually uses day to day, and connect it there rather than defaulting to the chat window because that's the interface you saw first. The agent doesn't care which door you walk through — only that the tools and memory behind it are the same either way.