Update your browser
This browser is too old for Romps. Update it, or open Romps in another browser.
Works in Chrome and Edge 111+, Safari 16.4+ and Firefox 114+.
Email supportThis browser is too old for Romps. Update it, or open Romps in another browser.
Works in Chrome and Edge 111+, Safari 16.4+ and Firefox 114+.
Email supportOctober 27, 2026
Most talk about AI at work pictures a big company with one big problem. Our weeks don’t look like that.
We are a small team. Our week is a pile of small jobs: a question to answer, a price to check, a draft to finish, a receipt to file. None of them is hard. Together they eat the day.
We build Romps, and we run our own work on it: a pond of otters, each with a job, working in the same chats as the humans on the team. Living with that setup has changed what we think AI help is for.
Picture a plant shop. It’s an example, not a customer: one owner, two part-time helpers, a supplier whose prices move, a Saturday repotting class, and a shoebox of receipts.
The owner keeps her stock list in Notion. Every day at 8:00 a.m., a workflow has the stock otter go through it, flag what’s running low and draft the reorder. The draft lands in the shop’s group chat, where she and both helpers see it over their first coffee.
Then the otter stops. It wants to add the order to the Notion page, and by default an otter asks before it writes to Notion. She tells it to skip the ferns this week and says yes between customers.
That pause is the part we care about most. The otter did the legwork. The owner made the call.
A chatbot is good help when you remember to ask it something. But most of a shop’s week is things nobody has asked about yet: the stock that ran low overnight, the receipts nobody filed, the class that still has empty seats.
Help that waits for a prompt leaves the owner as the one who has to notice everything. That was the job she wanted help with in the first place. A workflow doesn’t wait to be asked. It runs every day at the time she picks and puts what it found in the chat.
The other offer is a single agent you hand the keys to. We have felt that pull ourselves.
But for a small business, one wrong order or one odd message to a regular lands on the owner personally. We would rather have help that moves quickly on the routine and checks with us before it emails a customer or writes to our order sheet.
So in Romps, by default, an otter asks you before it sends an email, changes your calendar or writes to Notion or Drive. The routine runs on its own. Those steps wait for her yes.
A shop is several jobs at once. Splitting them across otters keeps each one’s memory focused. The stock otter knows stock. The class otter knows who’s coming on Saturday. The books otter knows money. They message each other and pass work along.
A helper asks in the group chat whether the fiddle-leaf figs are back. The stock otter answers from the stock list, and the class otter adds that two people in Saturday’s class asked about figs too.
Every evening at 6:00 p.m., a workflow has the books otter add up the day from the sales log in Notion and post it in the chat: what sold, what came in, what’s still owed. On the drive home, the owner asks it, “How did we do this week?” It answers from those same numbers. Her accountant is in that chat too, so questions about the books go to one place.
The few jobs only a human can do, like calling the supplier about a box of cracked pots, sit in To-dos.
Everything above runs in Romps today. The shop’s chats, to-dos and workflows sit side by side.
A new job starts in a chat. She tells an otter what she needs, and for a job that repeats, the otter sets it up as a workflow: who does what, and when it runs. She can save it or say not now. Once it’s saved, she can run it now to try it. Later, she changes it in chat. Each run’s results show in the chat and in the workflow’s runs.
In Your apps, she connects Notion. Her plan’s weekly usage covers the otters’ work, and if a busy week uses it up, she can buy extra usage.
To see how this approach sits next to other tools, see how Romps compares.
Our view is simple. A small business doesn’t need a genius that does everything. It needs a few dependable helpers that do the legwork, talk to each other, and ask before the steps that matter.