Safety, data and policy
Powerful agents. Sensible limits.
Plain-English guides for businesses giving AI agents access to email, files and systems: what to allow, what to approve, how to test and how to write it down as a policy your team can follow.
Guides
Agent permissions and approvals: give AI only the access it needs
How to decide what an AI agent may read, what it may do, and which actions a person must approve first.
Read the guide GuideKeeping data safe when you use AI tools
What not to paste into AI tools, what to check in a provider’s terms, and how to stop outside content tricking an agent into leaking information.
Read the guide GuideHow to test an AI agent before you trust it
A simple, repeatable way to check an agent on your real work, including the awkward and malicious cases, before it touches customers or money.
Read the guide GuideAn AI and agents policy for your business: template and checklist
A short, plain-English policy you can adapt in an afternoon, plus a checklist to run before any new AI tool or agent goes live.
Read the guideUse cases
Use cases where these controls matter most
Clear your inbox with a ten-minute daily triage
Let AI sort the day’s email into reply, delegate, read later and ignore, and draft the easy replies, with you pressing send on every one.
Read the playbook Customer supportPut an AI assistant on your website for common questions
An AI chat assistant answers routine questions from your help content around the clock and hands anything tricky to a person.
Read the playbook Admin and productivityAutomate repetitive work between your apps
Connect your forms, inbox, spreadsheet and CRM so routine steps happen on their own, with AI handling the reading and sorting.
Read the playbookNewsletter
One useful AI use case a week.
A short email with a practical use case, the tools, a ready-to-use prompt and how to tell whether it pays back. No hype.