Okou is an AI teammate for everyday business work. It connects to the tools your team already uses, brings together the information a task needs, and produces work you can review and share.
Tell it the outcome you want, like "review the customer history, prepare a brief for our meeting, and draft the follow-up." Okou works out the steps and uses your connected tools to carry them through.
Why we're building Okou
We began building an AI teammate in November 2025, after seeing the potential of combining natural language automation with cloud infrastructure. We were guided by a simple belief: working with an agent should feel natural. We still see a wide gap between the experience people need and what today's agent products deliver. We aren't building Okou to add another name to the list or follow the latest wave. We're building Okou to close that gap and make powerful AI approachable in everyday business work.
For a founder, marketer, salesperson, or operator, that means starting with a goal in plain language and getting something useful back. You shouldn't have to become an AI expert, memorize special commands, or wire together a flowchart just to get help with your work. Making that experience clear and natural is the standard we're building toward.
Three things matter especially: acting on context, communicating in the right medium, and rethinking how teams work together in the age of AI.
Act on context
Useful work starts with understanding the situation. Okou brings together internal context — customer conversations, team decisions, documents, and the current state of a deal — with external data, such as company research, market information, and relevant public sources. It can check changing facts against their sources and identify gaps before acting.
Connections to tools such as Slack, Gmail, Notion, Drive, and your CRM let Okou use that information to prepare a report, draft a follow-up, or update an authorized record. Custom connectors extend this to your team's internal systems. Your goal, previous decisions, and constraints help shape what it does next.
As you refine the work in the same chat, earlier messages, results, and corrections remain part of its context. You can explicitly reference another chat when its background is relevant. The work runs in the cloud, so it can continue after you close your laptop. Several tasks can run at once, allowing a long research job and a quick request to progress independently.
Before a customer meeting: Ask Okou to combine your CRM history and customer emails with recent public information about the company. It can prepare a brief covering what has already been discussed, what has changed, and which questions still need answers.
Today, automations let agreed work run on a schedule or a supported event trigger, such as an incoming email. Our ambition is to make working with Okou increasingly proactive: helping you notice what needs attention and move work forward with less prompting, within the boundaries you set.
Communicate in the right medium
Getting work done includes making it easy for someone else to understand. Okou can turn research, data, and decisions into a form that fits the audience and the next step:
- A message for a quick update or a clear next action.
- A webpage for a report people can explore and return to.
- A document for details that need review, comments, and editing.
- A presentation for walking a group through a recommendation.
The medium should help people see what matters, understand the reasoning, and decide what to do next. As soon as work involves more than one person, this becomes essential: teammates need to understand the result and carry it forward with fewer repeated explanations and rounds of clarification.
After the customer meeting: Give Okou your meeting notes and ask it to turn the discussion into a proposal document for your team to review, plus a concise follow-up email draft for the customer. The document holds the detail; the email makes the agreed next steps easy to see. You review the work before sharing it.
More done. Same team.
We're rethinking how teams work together in the age of AI. People and AI teammates should be able to work toward a shared goal, with less time spent re-explaining context and coordinating every handoff. One person's work should give the next person a useful starting point, and a better way of working should become something the whole team can use. That is how everyone on the team can level up and get more done together.
Shared agents give teammates a common role and instructions to work with. Shared workflows capture a reviewed procedure, reference material, and output standard. Each person runs the method with their own connected accounts and permissions, and can attach their own schedules or supported event triggers through automations.
For the next customer: Once you have reviewed and refined the briefing and follow-up process, save it as a shared workflow. A colleague can use the same method for another account, building on what your team has already learned. One person's good way of working becomes a capability others can use and improve. See Okou for teams for how sharing works.
Work from the channels you already use
You can reach Okou through the web app and the messaging channels enabled for your workspace, including Slack, Telegram, and phone messaging. The same agents and connected services can be available across these entry points, so you can hand off work from a familiar place.
Conversation context belongs to each chat thread. Continue in the same thread when you need its history; a new thread starts a separate context unless the channel explicitly opens an existing chat.
Choose your model and stay in control
You can choose a model for a chat across the Claude and GPT families, with personal and workspace defaults. Where supported, you can use your own Claude or Codex subscription, including on Free. See Models for the available options.
Access is controlled per connector and action, per agent. For example, you can allow Okou to prepare an email draft without granting permission to send it. Actions that require approval pause for your confirmation. Credentials stay on the platform, outside the model's reach, and the execution log shows which account was used, what was read, and what changed. See Permissions for details.
How a request becomes a result
- Describe the outcome. Message Okou with your goal and any relevant files or background, or let an automation start an agreed task.
- Okou gathers context and does the work. It selects the tools and steps, asks when a material detail is missing, and works within the access you have granted.
- Review and build on the result. You receive the work and its execution log. Refine it in the same chat, share it when ready, and save a useful method as a workflow.
Where to go next
- Quickstart: your first useful task.
- What Okou delivers: the range of work Okou handles, with prompt examples.
- Example workflows: complete examples you can adapt to your work.
- Workflows and Automations: turn a useful method into work you can repeat.
