Guide
Questions community managers ask
These are the six we are asked most. The short answers are on the product page; these are the same six with the reasoning left in.
Will members find it creepy?
The honest answer is that they would, if it messaged them about things they had not agreed to. So it cannot. Every introduction is double opt-in: the agent puts it to both people, with the reason each of them would care about, and nobody is connected until both have said yes. Neither person learns that the other declined.
Members set their own contact preferences, including not being contacted at all, and can change them in the same chat at any time. Cadence limits sit on top of that and cap how often any one member hears from the agent across every kind of message, not one flow at a time, so a quiet week stays quiet even when several things fire at once. And a human on your team is always one message away: anything the agent should not handle is passed to you with the thread attached.
Does it depend on WhatsApp?
No. It is omnichannel: WhatsApp Business, Slack, Telegram, email and voice today, with Teams next. Most communities start in the channel their members already use, often WhatsApp, but nothing in the product assumes it.
The channel account is yours and stays in your name, so you can move, add a second channel or leave without asking us. What we build is the layer underneath: the member graph, the workflows and the record of what has been sent to whom. The channel is a pipe.
Is it just a wrapper around a chat model?
The model is one swappable part, and you supply it. The product is everything around it: 325 workflows, a scheduler with 39 job types that decides what should happen this hour, consent and cadence controls, a member graph that remembers what every member has actually said, and a year of data on what members reply to and what they ignore.
A model with no memory of your community, no record of who has been introduced to whom and no rule about when to stop is a drafting tool. Knowing which two members should meet this month, and whether now is a good moment to ask, is the part that takes machinery.
Will the big platforms build it?
A platform has to get your members inside its own app before it is worth anything to them, and most members will not go. That is the bet the last decade of community software made, and it is why engagement dies a few weeks after launch.
We work the other way round. We live in the chat your members already use, we never ask them to install anything, and the value we add is in what we know about your community rather than in owning the place it happens. Anyone who wants to compete on those terms has to give up the app.
What do we have to set up?
Your channel account, your model keys, and connections to the CRM and calendar your team already works in. That is the list.
We do the setting up with you, on the house, and we bring your existing members in from whatever you keep them in today. CommonOS runs alongside your infrastructure, not inside it, and the records stay in your systems.
Can we change how it behaves?
Every default is yours to change, and you change it by talking to the agent in plain language: who it contacts, how often, what it proposes, which members it should treat carefully, and what it must never do without asking you first.
Hundreds of workflows and thousands of small decisions come already made, because a community manager should not have to configure a system before it is any use. Accept all of them, or change any of them.