The old way of running a service business.
You open the day with seven tabs.
CRM. Invoicing tool. Time tracker. Bank account. Calendar. A chatbot. Spreadsheet. Each one knows a piece of your business. None of them know you.
You spend the first forty-five minutes of every day moving information between them. Typing the same client name into four apps. Copying invoice totals into the bookkeeping spreadsheet. Setting calendar reminders for follow-ups you'd already written down somewhere else.
You forget to invoice. You forget to follow up. You don't know if you made money this week until your accountant tells you next month.
And every "AI" feature you've turned on writes emails you wouldn't have sent. The chatbot in the corner of your CRM offers suggestions so generic you'd be embarrassed to send them. The "AI assistant" in your invoicing tool drafts copy you immediately rewrite. The tools say AI-powered on the marketing page and autocomplete in the product.
"Most AI sidebars are autocomplete with a marketing budget."
The new way.
You open one app. It tells you what happened yesterday and what matters today. The morning briefing knows your overdue tasks, your upcoming meetings, the deals that haven't moved in a week.
You type it, or just say it out loud:
Otto drafts four cards: the deal, the invoice, the calendar event and the timer. You glance at each, fix the typo in the address, and hit approve. The deal lands in the pipeline, the invoice in Anna's inbox, the kickoff on your calendar, and the timer starts running.
The work is done before your coffee is.
"Four cards, one approval, and the morning admin is finished."
Why this is happening now.
Three things changed in the last eighteen months, and none of them is going back.
One. Tool-calling LLMs got reliable. Modern frontier models can now drive structured actions inside an application well enough to ship to working operators. The conversation moved from "can the AI write an email?" to "can the AI move the deal, generate the invoice, and start the timer in one breath?" The answer is yes.
Two. The interface disappeared. Plain English became a real input format. Operators stopped wanting to learn another dropdown menu and started expecting to describe the outcome and have software figure out the form fields.
Three. Operators are done with tab sprawl. The 7-tool stack is no longer acceptable. The market has moved from "can you do this?" to "can I do this without thinking?"
"Operators want to describe the outcome and let the software fill in the form."
Why the incumbents can't follow.
HubSpot, Salesforce and Pipedrive have CRM-shaped data models. They cannot ship voice-driven invoicing, time tracking, or bookkeeping without rewriting half the product, so expect them to bolt on a chatbot and stop there.
HoneyBook and Dubsado use AI to fill in templates, and their roadmaps keep adding workflow steps. You'll get prettier proposals from them, but an agent that does the work needs a different kind of software underneath.
ClickUp, monday and Notion are built around projects, with no real invoicing or money flow. They'll show you a kanban board with an "AI" badge that still can't send an invoice.
The category is open because the players who should own it are the wrong shape to build it.
What we're building.
Hey Otto is an AI operations platform for service businesses. One app holds the pipeline, invoices, time and money. You describe what you need in plain English and Otto hands back editable cards to approve. Thirty-five tool-calling actions, with 30-day undo on every one.
And we named him Otto because otters are one of the few animals that use tools, and the whole product is a tool-calling agent.
We are building it for the operator. The freelancer who runs her business at 9pm on a Tuesday. The four-person studio in Tel Aviv billing in three currencies. The dev shop in Buenos Aires that tried HubSpot and bounced. The consultant in London who has tried every "AI feature" and is tired of the chat box that drafts emails.
Those are the people Otto works for.
"Otto is an AI operator who runs the back office of a service business."
If that sounds like the tool you've been waiting for, you're who this is for.
The founder
