Our direction, while Bubbling AI is in development. The product is not open for sign-up yet.
We are building Bubbling AI with people in the picture. Human judgment, taste, experience, and curiosity are part of the value we want to support. Our product direction is to help people do more while keeping a meaningful place for their contribution.
We are not trying to develop you out of the picture. That is our bet. If the industry takes a different direction, this is still the kind of company and product we want to build.
A faster answer is useful. So is a conversation that helps you notice the right question, test an idea, or decide what matters. We want Bubbling AI to support both: thinking together and moving the work forward.
The person is part of the value
People care about different parts of their work. One person wants to shape the design. Another wants to understand every implementation detail. Someone else wants help with the logistics so they can concentrate on the argument, the customer, or the difficult decision.
Our aim is to make those differences a starting point. You should be able to contribute where you want to contribute and hand over work where assistance is useful. Success should include whether the result reflects your priorities, whether you understand the important choices, and whether the process leaves you more able to do what matters to you.
This is the future we want to help build: people and AI working together, with people’s contribution valued as the technology becomes more capable.
Discuss the route, keep the direction yours
A useful thinking partner does more than accept an instruction and disappear. It helps establish what you mean, what a good result would look like, and which choices need your involvement.
The experience we’re working toward is simple to describe: you choose the destination, we discuss the important parts of the route, and the assistant helps carry out the work. Routine steps need not become a stream of questions. Material changes in direction should not happen silently.
Learning your preferences should make that collaboration more comfortable. It should not turn yesterday’s preference into permission for every future action. You need ways to correct the assistant, change your mind, and set a different direction.
A company can benefit without treating every thought as shared knowledge
We believe trust is part of the business case for AI. It is reasonable for a person to hesitate before bringing an unfinished idea to a system when they cannot tell who will read it or how it will be used.
A shared report and the private conversation that helped produce it serve different purposes. A team needs access to its deliverables. A person also needs room to ask a basic question, reconsider an assumption, and explore an idea that might go nowhere.
Our company-workspace design distinguishes Personal to each member from Held by the company. In Personal mode, work memory and private chats are protected from ordinary company-admin access. Held mode permits company access. Documents, uploads, and conversations deliberately shared into a project are treated as company content. These are product access rules, not a determination of legal ownership.
The distinction must be visible before someone joins. Changes need to explain what happens to existing information as well as future work. Exceptional access and retention also need their own clear explanation; they should not be hidden behind a reassuring slogan. The full policy will be published before company accounts open.
We cannot promise anyone continued employment or control how a business reorganizes. What we can do is choose the kind of product we are building: one that respects people’s contribution and makes its data boundaries understandable.
A conviction, expressed through the product
A private place to explore an idea, a useful discussion before execution, a visible decision that needs you, and a way to carry your knowledge forward are all expressions of the same choice. We want the design to support a future in which people and AI work together. It is a conviction about what we are building, not a prediction that technology cannot replace tasks or a guarantee about anyone’s employment.
The work ahead
A statement of values is only a beginning. The interface, permissions, data policies, and everyday behavior must support it. We will explain those choices as the product takes shape, including the limits and the places where work remains.
We want the result to feel like a place where you can think clearly, do useful work, and remain involved in the decisions that matter. That is the standard we want people to judge us by.
Read about how better models should improve the workspace, portable memory, and our approach to privacy.