Run AI workflows without giving up control.
Apps are connected as personal or shared connections, and you choose which teams can use them. Important actions pause for a person. Approved actions are re-checked before they run. Every run leaves details you can read.
Ten answers, in the order you would ask them
What data can willdo access?
Only what a team has been granted through connected apps. willdo reads through those apps, acts only through app actions that carry a permission, and each organisation's data lives in its own database schema. Apps a team has not been granted do not exist for it.
How are app connections stored?
App connected accounts and model keys are stored encrypted and managed at the organisation level. You choose whether a connection is personal or shared, and you can reconnect or remove it from one place.
What can an AI workflow do automatically?
Reads and drafts. Reading a ticket, reading a charge, drafting a reply or a brief run without asking and are recorded in the run details. Drafts are saved as documents, not sent.
Which actions can require approval?
Sending a message or email, creating or changing a record, and any external action ask first by default. Financial effects, deleting data, and changing access never have an Allowed option, whatever approval mode you choose. Every person sets Allowed, Ask first, or Blocked for themselves. Approval requests expire after 30 days and decline by default.
Which AI model providers may process data?
The hosted AI included with willdo, or a provider account you connect: Anthropic, Google, OpenAI, xAI, Mistral, DeepSeek, Cloudflare, or a self-hosted OpenAI-compatible endpoint. You set an organisation default and can override the model per lead agent or specialist agent.
Is customer data used for training?
Model providers process the prompts a run sends them. Retention and training depend on the provider and the account you connect; on your own keys, your provider's terms apply. Ask us for the current hosted-provider terms at [email protected].
How can data be deleted?
Remove a connected app from your organisation to stop access immediately. To delete stored run history, documents, or your organisation, email [email protected] from an Owner account.
How are incidents handled?
Report anything unexpected to [email protected]. A written incident-response summary is available on request while the public policy page is being finalised.
Which subprocessors are used?
The AI model provider you choose, plus the hosting and email providers listed in the Data Processing Agreement. Request the current list at [email protected].
How can security concerns be reported?
Email [email protected] with “Security” in the subject. Please do not include customer data in the report itself.
Governed at the organisation level, not per prompt
What you would expect
Email verification, Google sign-in, forgot password, and TOTP two-factor with a recovery key.
Who can do what
Owner, Team Lead, Manager, and Operator roles. Every role can be bound to specific teams.
Details for every run
Steps, results, decisions, action receipts, cost, and a final status. Business view by default, developer view on demand.
What willdo is — and isn't
willdo is
- the platform your business runs work on: rules, an AI team, and human approvals across your apps and playbooks, with the details of every run
willdo is not
- a replacement for legal, compliance, or security review
- a reason to give AI blanket app access
- a fully autonomous employee
- a guarantee of compliance, or that every AI output is correct
Grant one team one app, and approve the first request yourself
Then read the run details. Questions: [email protected].