Amy builds Open Amy

Meet Amy. Put AI to work for your team.

Lead the AI
transformation.

AI does the work. On your terms.

Your people explain how work gets done. Amy turns their knowledge into shared AI instructions, guided processes and apps that run without coding or deployment.

Bring your team into an Agentic Workspace: a shared place to work with AI, with defined access, chosen AI resources and visibility into activity and costs.

Start with your work Explore your controls

From the way your people work to the way your company uses AI.

Meet Amy: your AI architect, with glasses, a dark hoodie and a friendly smile.
Team daysIllustrative example

AI handles the request.
Your manager decides.

  1. The request

    Maya: “I’d like three days off.”

  2. Your rule Active rule

    More than 2 days off requires manager approval.

  3. Human approval

    Waiting for a manager

  4. The record

    1. Request submitted Pending · 3 days off

    2. No approval recorded yetThe request stays pending.

Your people know the work. Your rules guide the AI.Find a place to start

First, capture what your people know

AI needs the details
your people take for granted.

Who decides? What makes a case unusual? When should someone step in? Your team explains the work in their own words. Amy asks about the details and turns them into a connected description you can inspect and improve. That is your intent.

  1. Explain it. Amy draws it.

    The person who knows the work can explain it without drawing a diagram. Amy connects the people, steps and rules in trees and charts the team can review.

  2. Find the missing decisions

    Amy checks how the parts connect, helping you spot gaps, conflicting rules and unclear handoffs. Work through the exceptions before handing the task to AI.

  3. Agree what should happen

    Review and refine the intent with the people who own the work. Keep it as a shared process model, use it to guide AI, or build on it later. Modelling alone is a useful result.

A question worth asking, for example“What makes a request urgent — and who decides?” Amy helps make the meaning and the responsibilities explicit.

From a conversation to shared understanding

Your team explains.
Amy asks the next question.

“We always check with the manager” sounds simple. Amy helps make the important details explicit: when approval is needed, who gives it, and what happens while a request waits.

The conversation becomes an intent your team can review, share and refine. Amy keeps the meaning and its connections together. Follow the full journey

A conversation with AmyIllustrated example
You

I want to make requesting time off easier.

Amy

Who should approve a request?

You

Their manager.

Your intent takes shape

Goal
Make requesting time off simple.
People
Team members and their managers.
Rule
Time off is confirmed after the manager approves it.
That is your intent. You can review it, use it, share it and keep refining it.

Your Agentic Workspace. Your rules.

Who has access?
Which AI? Whose decision?

These questions belong in your AI rollout from the start. Bring your people into a workspace with access to the work they need, backed by model policies and visibility into usage.

Model choices

“Which AI are our people using?”

Make the provider and model visible. Deployment administrators allow or block models for modelling and app use. Workspace owners manage the shared models available to their team.

People & access

“Who can see and do what?”

Invite people into the workspace and assign access to the apps and processes they use. Keep authoring collaboration for apps, processes and prompts separate from permission to use a running app.

Permitted actions

“Will AI make the decision for us?”

Define the actions, responsibilities and approval steps in the intent. Assign people their app and process roles. Amy checks access and permitted data changes when the work runs.

Activity & spending

“Can we see the activity and cost?”

Review recorded app activity and process history. Deployment administrators can inspect model calls by user and model, token usage and estimated costs, so AI use is easier to account for.

People and AI, working together

Your Agentic Workspace

A shared place for your team to use AI-powered apps, follow guided processes and work from the same agreed rules. People find their tasks, connect with colleagues and act within the access you define.

Explore the Agentic Workspace

Put the understanding to work

Apps that act.
People who stay involved.

An agentic app uses AI to help carry out the work: gather information, act on a request, or bring a decision to the right person. The screens and actions adapt to the task, guided by your rules.

Amy runs the intent you reviewed directly. Your team maintains the rules, without a separate application codebase to build or deploy.

For work that moves between people: a guided process takes each case through its steps, decisions and handoffs, keeping people involved where needed.

  1. Explain the work
  2. Review the rules
  3. Put AI to work
See how actions and approvals work
Holiday Planner running in Amy, with places to explore, suggested actions and space to ask for something else.
Inside Amy · Holiday Planner. An agentic app, running from its intent.

Start with a real piece of work

One team. One process.
A useful first step.

Capture the knowledge, guide a process, or put an AI app to work. Start where your team needs help today.

Understand / Process model

Capture the knowledge

Make the way your team works visible. Review responsibilities and exceptions together before deciding what to change or delegate to AI.

Start here“Our onboarding only works because one person knows every step.”
See how Amy models it
Guide / Process

Keep a case moving

Amy guides people through the work, case by case. Make decisions, responsibilities and the next handoff clear.

Start here“Requests keep getting stuck between our teams.”
Explore guided processes
Act / Agentic app

Put AI to work

Turn agreed rules into an app your people can use. AI helps carry out the task, with access and approval steps defined for the work.

Start here“Let AI handle requests, with a manager approving longer leave.”
Try the approval example
Reuse / Prompt

Share how AI should work

Give your team a reusable prompt with shared context, rules and a clear description of a good result. Review and refine the instructions together, then use them with your chosen AI.

Start here“Everyone gives our AI assistant different instructions.”
Shape your instructions

Need a website, web application or mobile app in your own stack? A developer can build from the same reviewed intent.

When the rules change

Keep AI working from
the rules you mean today.

A new approval step. A different owner. An exception nobody had written down. As instructions grow and get copied, details can drift or conflict.

Amy connects your rules to the people and steps they affect. Review the proposed changes together and update the intent that guides the work.

“Longer holidays now need HR approval too.”

One new rule. Two parts of the intent to revisit.

The time-off intent, before and after

Illustrated update
01

Who is involved

Team member → Manager

02

What happens

Request time off → Manager approves → Time off confirmed

The manager approves the request before time off is confirmed.

Amy

For teams and the people helping them adopt AI

Which piece of work
would you hand to AI?

Bring one process and the person who knows it. Work with Amy to make the steps, decisions and boundaries clear, then choose how to put them to use.

Explain your work to Amy See plans and pricing