From human knowledge to AI at work

Your people know the work.
Start with them.

Start in Studio with one process and the person who understands it. Amy helps make the steps, decisions and exceptions explicit. Review them together, then bring your team into Space to put the work to use.

01 · Explain

A conversation is enough to begin.

The knowledge might belong to a colleague, a process owner or a client. They describe what happens in their own words. Amy asks about the details that matter.

No diagrams to draw. No code to write.

Amy connects the explanation into an intent: a structured description of what should happen, who is involved and which rules apply.

  • Start with the everyday task and the result you need.
  • Explain exceptions, judgment calls and approval steps.
  • See the work take shape in trees and charts.
Amy’s conversation beside the intent tree, making the explanation visible for review.

02 · Understand and review

Find the gaps before they become someone’s instructions.

Amy checks connections and helps uncover conflicting rules, missing decisions and unclear handoffs. The people who know the work review what it should mean.

Try it · Expense Approval template

Follow the work. Inspect the decision.

Open full page

Select Review to see what the manager checks and where approval or rejection leads. Switch to Tree or Linear+ to explore the same model from another angle. No sign-up needed.

See the work

People, steps and decisions

Read the tree or explore a process chart. Follow who does what and where the work goes next.

Work through exceptions

“What happens in this case?”

Clarify the unusual requests and the decisions that need a person. Logic checks help focus the review.

Keep a shared reference

Review the changes

Inspect Amy’s proposals before accepting them. Versions help you revisit how the intent changed.

A shared process model is a complete result. Use it to agree how the work should happen, even if you never automate it. Explore all three templates

03 · Give AI shared instructions

Your team’s expertise, ready to reuse.

For a Prompt project, Amy helps define the context, inputs, boundaries and expected output. Share and refine those instructions with collaborators, then take the finished prompt to your chosen AI.

Illustrated instructions · Service requests

Make a good result explicit.

“Assess the request against our service criteria. Identify missing information. Recommend a priority and explain why. Flag exceptions for the team lead. Return the assessment, supporting facts and points needing review.”

Review the outcome against the intent.

Give people a common reference for what the AI should produce and where human review belongs. Refine the prompt when the policy changes. When used in another AI tool, execution and access follow that tool’s controls.

04 · Bring your people in

Your team’s Space.

A shared place where your team works with AI—using your apps, following your processes, and working within the permissions and AI access you define. The relevant context and agreed rules stay connected to the work.

Your work

Apps, cases and the next task

Find the apps and processes assigned to you. See which cases need your role, what is waiting and what has been completed.

Your team

People and conversations

Find colleagues in the team directory and talk through the work in private, one-to-one messages.

Your controls

Defined access and AI resources

Workspace owners assign app and process roles and manage shared AI model connections. Deployment administrators set model policies and review usage and estimated costs.

Owners manage invitations and settings in Studio. People can use Space without permission to change the underlying intents.

Explore Space

05 · Put it to work

Your rules become the basis for the work.

Choose an app for interactive work or a guided process for cases and handoffs. Configure the AI resources and access, try the behavior, and bring in the people who will use it.

No separate app to code or deploy.

Amy runs supported behavior directly from the intent. The team can refine its rules without maintaining a separate application codebase.

Human decisions stay explicit in the process. App actions pass through permission and data checks before changes are stored.

See how Amy runs the work
An app running in Amy, showing a record and the actions available for it.

Start with something your team knows

Explain one piece of work.
Decide what AI should take on.