“It launched in the wrong account.”
An agent that does not know which client it is on can write to the wrong ad account. Win & Scale tells it the workspace and ad account first, and asks for that call before any write.
Use casesAgencies
One AI agent for every client’s Meta ads.
Run every client from one agent. Each ad account has its own workspace, so credentials, creatives and launches stay with their client.
7 days free with every feature. Cancel anytime.
Try asking
- Which workspace are we on?
- Switch to Doorlark and check its evidence health.
- Which workspaces have no ad account connected?
What goes wrong when one agent serves many clients?
The risk is not the agent’s skill. It is not knowing which client it is working on.
“Client tokens live in a shared doc.”
Every client’s Meta token ends up pasted somewhere shared. Win & Scale keeps each workspace’s credentials encrypted, and the settings page shows only a masked summary.
“One client’s lesson showed up in another’s brief.”
Values, learnings, briefs and strategy belong to one workspace. A lesson written for one client stays in that client’s workspace.
Each client is a workspace with its own ad account.
One workspace holds one client’s Meta ad account and its own encrypted credentials. An admin connects the account once, and everyone else works inside it.
- Admins connect the Meta account, name the workspace and manage the roster.
- Members launch, sync and read reports.
- Invite a client or a teammate by email. Nobody sees a workspace they are not a member of.
Your agent always knows which client it is on.
Your agent asks which workspace the session is on before it writes anything, and moves to another one with a single call.
- meta_get_workspace returns the workspace, your role and the ad account every read and write will hit.
- meta_list_workspaces lists every client you belong to and marks the active one.
- meta_switch_workspace moves the session to another client. It touches nothing on Meta.
- An rsk_ API key for an unattended agent is pinned to one workspace. It sees only that client, and meta_switch_workspace refuses it.
Every launch lands paused, in the workspace you chose.
Your agent builds the campaign for the active client with nothing spending. Activating is a separate call, and your agent asks you in the conversation first.
- meta_launch_ads creates the campaign, ad set and ads paused, in the active client’s ad account.
- meta_list_presets shows each client’s own launch preset, so a launch starts from their defaults.
- meta_activate is the one call that spends.
What does each client cost?
Each client’s ad account pays its own flat monthly price, set by that account’s ad spend. Every tier includes every feature.
Ad spend up to $10k a month
$49/ month per ad account
7-day free trial, every feature
Start freeAd spend up to $25k a month
$99/ month per ad account
7-day free trial, every feature
Start freeAd spend up to $50k a month
$199/ month per ad account
7-day free trial, every feature
Start free
- Start free
Ad spend up to $100k a month
$299/ month
- Start free
Ad spend up to $250k a month
$399/ month
- Talk to us
Ad spend over $250k a month
Custom
Worked example
Three clients, monthly list prices
A client at $8k spend
$49 a month
A client at $20k spend
$99 a month
A client at $60k spend
$299 a month
What can you ask your AI agent across clients?
Ask the way you would ask a media buyer who runs your whole book. Your agent checks the workspace before it acts.
- Where are weWhich workspace are we on, and which ad account does it hit?
- Switch clientAnswered in the conversation: Switch to Doorlark and check its evidence health.
- Missing connectionWhich of my workspaces have no ad account connected?
- Round the bookGo through each workspace and list the tests that are ready to call.
- Launch pausedSwitch to Saltmilk, confirm the workspace, then launch the three approved cuts paused.
- Hand overSummarise what we learned for Vessa Row this month.
Claude Code · winandscale connected
Switch to Doorlark and check its evidence health.
meta_get_workspaceVessa Row · adminmeta_list_workspaces3 workspacesmeta_switch_workspaceDoorlarkSwitchedmeta_get_evidence_healthsynced 3h ago
I’m on Doorlark now, and its ad account synced 3 hours ago. I changed nothing on Meta. Your browser is still on Vessa Row, so pick Doorlark there too, or the next page load moves me back.
Questions, answered
Everything else is in the docs.
- Can an agency use Win & Scale for several clients?
- Yes, an agency gives each client a workspace, and each workspace holds one Meta ad account with its own credentials. Your AI agent moves between clients with meta_switch_workspace.
- How does my agent know which client it is working on?
- Your agent calls meta_get_workspace, which returns the workspace and the ad account every read and write will hit. Its description asks agents to call it first and before any write.
- Are one client’s Meta credentials visible to another?
- No, Win & Scale stores each workspace’s Meta credentials encrypted, and only the workspace’s admins connect them. The settings page shows a masked summary, never the token.
- Can a headless agent work across several clients?
- No, an rsk_ API key is pinned to one workspace, and meta_switch_workspace refuses it. Use one key per client, or a browser sign-in for an agent that serves several.
- If my agent switches workspace, does my browser follow?
- Not fully. A browser still open on another workspace moves the agent back when a page loads, so pick the same client in the app as well.
- Who can change a client’s Meta connection?
- Only an admin of that workspace can change the Meta connection, the workspace name and the roster. Members can launch, sync and read reports.
- How much does Win & Scale cost for an agency?
- Each client’s ad account pays a flat monthly price set by its ad spend, starting at $49 a month for up to $10k of monthly spend. Every ad account starts with a 7-day free trial.