Early access

Security and control

Your clients’ accounts stay in your hands.

AI does the legwork. HeyFlory makes sure it only sees what you allow, never touches a campaign without a person’s approval, and keeps a record of everything.

Nine safeguards, built in. Not settings you have to remember to turn on.

Official sign-in only

You connect Google Ads, Meta and the other platforms through their own sign-in screens. HeyFlory never sees or stores your passwords.

Only the accounts you pick

You choose each client account HeyFlory may use. Everything else in your manager account or business portfolio stays invisible.

Read-only by default

Every account starts read-only. Uploading ads or changing campaigns must be switched on per account, by an owner or admin.

A person approves every change

When AI prepares a change, it waits in HeyFlory. Someone with permission reviews the exact action and approves it after a recent sign-in.

Every agency walled off

Each agency’s data is separated at the database level, so one workspace can never read another’s accounts, reports or alerts.

Credentials stay encrypted

Platform access tokens are stored encrypted and are only used by HeyFlory’s servers. They are never passed to Claude or ChatGPT.

Separate consent for each AI tool

Connecting Claude or ChatGPT is its own permission, per person. Widening what an assistant may do needs a new approval, never a silent upgrade.

A record of every action

Who asked, who approved, what was sent to the ad platform and what came back. Approvals are kept as a permanent record.

Disconnect in one step

Remove an assistant, an account or a whole platform at any time. Access stops straight away and pending changes expire.

How a change actually happens. Five steps, and a person is always one of them.

  1. Claude or ChatGPTProposes a changeFor example: pause “Barrier cream – Broad”.
  2. HeyFloryChecks permissionIs this account allowed to change, and is this person allowed to ask?
  3. Your teamApproves in the browserThe exact action, after a recent sign-in. Unapproved requests expire.
  4. HeyFloryApplies it onceSent to the ad platform a single time, never repeated by accident.
  5. Audit logKeeps the recordRequest, approval and the platform’s response, saved together.

What the AI sees. And what it never does.

Claude and ChatGPT only ever talk to HeyFlory. HeyFlory decides what each question is allowed to read, from the accounts you chose.

  • ✓Performance data for the accounts you selected
  • ✓Campaign, ad set and ad names and settings
  • ✓Only what a question needs, when it is asked
  • ✕Your ad platform passwords or access tokens
  • ✕Accounts you didn’t select
  • ✕Other agencies’ data, ever

Security questions. Straight answers.

Does HeyFlory change anything without asking?

No. Accounts are read-only unless an owner or admin enables specific actions for that account, and every change still needs a person’s approval in HeyFlory before it is sent.

Can Claude or ChatGPT see our login details?

No. The AI tool talks to HeyFlory, and HeyFlory talks to the ad platform. Access tokens stay encrypted on HeyFlory’s side and are never shared with the assistant.

Who on my team can see which client?

Owners and admins decide. Each teammate only sees the client accounts they’ve been given, in the dashboard and in their AI tools.

Do you have SOC 2 or ISO 27001?

Not yet. HeyFlory is in early access. We’re happy to answer your security questionnaire and walk you through how access, approvals and data separation work: write to hello@heyflory.com.

How do I report a security concern?

Email hello@heyflory.com. A person reads every message and replies directly.

What happens if we disconnect?

Access stops immediately, pending changes expire, and your AI tools can no longer reach that account through HeyFlory.

AI that helps, on your terms.

Read-only by default. A person approves every change. Disconnect anytime.