Agency

Adding & managing clients

Create a client, configure their account, and run their reporting from the management page.

I'm using the platform as a
All help topics

Each client you manage is a separate, isolated account. Here's the full lifecycle, from adding one to running it day to day.

Adding a client

From Clients, choose Add client and fill in:

  • Name (required) and website (the website auto-fetches a favicon you can keep or replace with a custom logo upload).
  • A contact email (where billing and account emails go), country (for holiday-aware scheduling) and currency.
  • Go-live date (optional) - when this client's site or tracking went live. Before this date, the Brain treats traffic, conversion and spend swings as pre-launch noise rather than real movements, so it won't alarm about launch artefacts (like a "+1,798% sessions" jump on a site that only just went live). See Client settings.
  • Data refresh - how often their connected accounts are pulled into the dashboard: Daily (the default), Weekly on Mondays, or Manual, which updates only when someone presses Refresh now. You can change it later from the client's page. See Data refresh & cadence.

Save, and you land on the client's setup: confirm what we read from their website, set their main goal, choose who connects each tool, add the kickoff call notes and invite them. Every step can be skipped. See Setting up a new client.

Add client stays greyed out, with the reason beside it, until your workspace's required setup steps are done. During your first 7 days that is only a support email, because your AI runs on ours while the trial lasts; after that your Claude API key and Google AI key are required too. Without the keys a client's dashboard would fail on every AI feature, and without the email their contact buttons would reach the wrong inbox. Finish them from More, Settings, Setup - see The setup hub.

The Clients list

The list is grouped by who needs you, so the top of the page is always the work:

  • Needs attention - live clients that are genuinely blocked: lapsed billing, a stopped data feed (A feed has stopped, linking to Connection Health) or a churn risk that has reached At risk. Red is reserved for these.
  • Setting up, Connecting and Data flowing - clients still being onboarded, grouped by the stage they are at. Each shows its next step as a link: Setting up - Connect the first platform, Connecting - Run the first data refresh, Data flowing - Publish the first report or Data flowing - Invite a portal user. Until the client is live, its page also has a Client setup button, next to Settings, that reopens the client setup. Once all four are done the client is Live and leaves these groups. The stage is worked out from what the account has done, never ticked off by hand.
  • Pending - live clients that need a nudge: a report overdue, a draft storyline to review, payment required, or a churn score at Watch.
  • Up to date - healthy, nothing to do.

Archived clients sit in a folded Archived footer at the bottom and are left out of every count. A card's task count is open work only - accepted and in-progress tasks, never suggestions that were rejected or archived.

"Scheduled reports are not running"

An amber warning above the list means a scheduled report for one of your clients was due more than a day ago and has not been generated. This is on the platform's side, not something in your setup: the scheduler that runs every morning has not picked those schedules up, and running a report by hand does not clear the warning. The platform team can see the same fault on their side, and scheduled reports resume on their own once it is resolved. The warning names the clients waiting and when the next run is due. If a client needs their report before then, generate it from their management page and publish it as normal.

The management page

Open any client to reach their management page - the hub for that account, grouped into clear zones:

  • Pulse - a short State of play written for your team, a digest of recent client activity, and Brain questions - things the Brain is unsure about, to answer or dismiss.
  • Reports - run analysis on demand and control which auto-generated weekly insights are published.
  • Goals & work - open the client's Storyline tab to set goals (which also decide the metrics on their dashboard's KPI Targets card) and publish storylines, and see their tasks.
  • Connections & usage - platform health for every integration, and AI token usage with per-feature caps.

The header shows platform health (e.g. "4/5 connected"), the last data refresh, a Refresh now button, and the client's setup stage with its next step. See The client Manage page.

Client settings

The client's settings/edit page is grouped into:

  • Profile - name, logo, contact email, who can sign in (by company domain or by individual email), country, currency, and the account tier (which controls feature access such as Creator's Studio).
  • Configuration - data refresh cadence and per-feature monthly usage limits.
  • Advanced - a link to the client's Brain ops page, plus Connected AI and Slack.
  • Danger zone - delete the client (admin only).

Connecting the client's data

Integrations are connected per client from their connection screen, using your agency-level manager accounts. The client grants your managed account access to their properties. See How data connections work and Connecting Google.

Running reports

Generate reports from the management page. Unlike self-serve accounts, agency reports are saved as drafts for your team to review, then published to the client when you're happy. See Generating reports.

Controlling what the client sees

You decide which dashboard tabs each client sees using the eye controls in the team bar on the client's own dashboard, and you can preview as the client to check their exact view before sharing anything. See Preview as client.