02 / Browser assistant
Moxby / Austin, Texas
Moxby offers a browser assistant for teams whose work continues on an active website. Discuss the help a person needs beside that page, with the browser environment and permission boundaries defined separately from the shared workspace.
A browser request is not a download or an installation. Availability and supported environments need confirmation before an offer is agreed.
The examples on this page describe requested behaviour. Photographs are illustrative workflow scenes, not verified product screenshots.
The offer / Work beside the page
Keep the shared task's purpose distinct from the website's current state.
The workspace can explain what the team wants done. The browser is where a person encounters the actual page, its fields and the permissions of the account they are using. Those are different kinds of context.
A browser workflow assistant inquiry begins with what the person needs at that point. Perhaps the request is to prepare information beside a form while leaving submission to the user. That is a narrower request than asking software to complete the whole task unattended.
Describe the activity first. The form of access, including any extension requirement, belongs in the availability discussion. This page does not offer an installable release or imply that a particular browser is supported.
01 / A browser-side example
Consider a user preparing a non-sensitive description for a website form. The team has agreed what the description should say. The user still needs to inspect the destination.
In this illustrative request, the assistant would help prepare the text while the person remains in the page context. The request ends before submission. It does not include signing in for the user or choosing a different destination if the expected form is unavailable.
This is a useful distinction for a setup inquiry because “help fill the form” can mean very different things. A proposed value shown for inspection is not a saved value. A saved draft is not necessarily a submitted record.
The task's context may explain the wording, but only the active website shows what fields are present now. If the site presents a different form, the original instruction may no longer be enough to decide what should happen.
Keep that uncertainty in the request. You can ask for a human review point without specifying how it would be implemented. The implementation and permissions need confirmation; the boundary can be clear from the beginning.
If the request includes actually changing a field or submitting a form, continue with Website actions. That page addresses the target and stopping condition in more detail rather than hiding them inside the assistant label.
Illustrative request / Description preparation
02 / Permission boundaries
Being able to open a website is not the same as authorising an assistant to act on it.
The site may allow a person to read material while restricting changes to particular roles. A browser or managed device may impose another set of limits. The inquiry should identify those constraints without asking anyone to weaken them to make a demonstration work.
Reading information and writing information also need separate consideration. A request to help interpret a visible page does not silently include editing it. A request to prepare a field value does not silently include submitting the form.
Do not send passwords or temporary access codes with a browser-assistant inquiry. They do not answer the permission question. Describe the type of access and any approval needed from the organisation that manages the account or device.
A permission prompt is not a security certification. The illustrative scene shows a decision boundary so it can be discussed; it does not promise a particular permission model or prove that all processing stays on the device.
A
Name the page material the requested help would need. An active tab may contain more information than the task requires. Do not treat access to unrelated content as an automatic part of a browser-assistant request.
B
Distinguish a suggestion from a field edit. If the assistant is expected to affect the website, describe the permitted target and what the user must inspect before the change takes effect.
C
State whether a managed environment or site owner needs to approve the proposed behaviour. A team member asking for help cannot, through an inquiry form, grant rights they do not hold.
D
Explain whether the request applies only while the user is actively working on the page. Leaving the tab or changing accounts should not be casually described as permission for the original request to continue elsewhere.
03 / Environment to confirm
“In the browser” describes a location, not a compatibility specification. Include enough about your environment to make the request assessable.
Give the browser name and version you use, if known. This identifies a requirement, not a supported platform. If several environments are involved, say which one matters for the first requested routine.
Explain whether the environment is managed and whether extensions or other software access require approval. Do not bypass a restriction to fit the example. An unavailable permission can change the feasibility of the requested setup.
Name the website or provide a public URL. Note whether the intended task is behind account access, without sharing a private link or account details. A publicly visible page and a signed-in work form may behave differently.
Say when the person is present and what they should check. If the requirement is actually for a job that runs without a person at the page, state that explicitly; it is a different scope and is not established by this offer.
If a chosen local document or external service supplies input, name that boundary in plain terms. The Local workflows explanation covers why local behaviour does not establish offline operation or a privacy guarantee.
Requested user action
The inquiry describes help while a user is working in a known page context. Preparation has an observable endpoint and the person checks the proposed result before authorising a further step.
Unattended execution
A request to run after the user leaves, to retry on its own or to choose a new target is not the same offer in different words. It needs separate feasibility and permission consideration; this page does not promise that behaviour.
04 / Back to shared context
The team needs to know where the browser work stopped.
In the description-preparation example, the result is text ready for the user to inspect. Calling it a completed website update would give the next person the wrong account of what happened.
If your routine needs an account of the browser activity in the shared task, describe what should be recorded and who should review it. Do not assume that a browser-side action automatically updates the workspace, or that every piece of page content should be copied back.
Keep the report proportionate. A task may need a statement that the preparation is awaiting review, not a copy of the entire page or details from the user's account. The intended relationship matters more than the volume of information moved.
The Team workspace offer covers the shared conversation and document relationships. Pair it with Browser assistant when your request genuinely needs both contexts, rather than treating the browser companion as the whole product.
05 / From request to scope
Start with the user's intended activity, such as preparing a description beside a work form. Explain what the person does now and what help they want. A public website name is enough to begin; the inquiry does not need private account access.
Say what the assistant should leave for the user to decide. If the request includes changing the site, include Website actions in the setup builder. If local inputs matter, include Local workflows and describe the environment rather than assuming access to it.
The browser and permission requirements need review before a proposed scope can be settled. This is where assumptions about extension access or managed devices belong. Naming a platform in the request is not proof that it is supported.
The option sheet produces a statement of interest, not an installation or an order. Individual scope and any commercial proposal follow confirmation of the requested environment and behaviour. Read How it fits for the full inquiry process.
Before making a request
If you need guaranteed support for every website or unattended execution without review, this page does not establish that fit. A clearly described browser-side activity is the right starting point.
No installable release is offered on this page. Describe the assistant behaviour and browser environment you need so availability and the appropriate form of access can be confirmed through an inquiry.
The illustration of a side panel is there to make the working context understandable. It is not an installation instruction or a claim about extension-store availability.
Browser compatibility is not established by this page. Include your browser and version, operating environment and any managed-device restrictions in the request. Support must be confirmed for the intended setup.
If a particular environment is a condition of your decision, make that clear before discussing other behaviour. Do not treat an example page layout as a compatibility test.
No. A request for assistance while a person is using a page is not a request for unattended execution. Any further behaviour needs its own scope and permission review.
For example, preparing text for inspection does not include automatically retrying a submission after an uncertain response. That would be another action with another consequence.
No. A local workflow can still involve a website or external service. Describe the input and output and identify any external connection; local does not establish offline operation or a guarantee about data processing.
Use the Local workflows page to describe that boundary. The privacy notice separately explains how this inquiry website handles the information you send.
Your browser-side request
Choose Browser assistant in the setup builder and describe the help you want beside the website. Add the workspace if the shared task is part of the request. The summary keeps those areas distinct so their boundaries can be discussed.
Local workflows and Website actions select Browser assistant too. That dependency makes the browser environment a visible scope question, not an extra product entitlement.