AI AGENTS · IN TESTING PHASE

Audit agents that turn a report into a decision

We're extending ClientDeck with AI agents that analyse accounts, rank priorities and prepare recommendations. We're currently piloting the agents with selected partners — once testing is done, we'll release them into the full version of ClientDeck.

Read-only client layer
Approval before execution
Execution rollback
ClientDeck AI audit and agent overview

What AI agents will deliver to clients once released from testing

Agents prepared by us

In the full version, the client picks an approved agent, picks a client and runs a read-only audit. During testing we run launches internally.

Recommended data up front

We show what to ideally connect: GA4, Search Console, ad accounts, feed or margins. The data layer is already in place.

Output in the client overview

We turn the agent's markdown report into an account score, priorities and an exportable output. After release it will be part of the client zone.

What we already have ready in the testing phase

Approval

Recommendations stay under team control. Execution doesn't run without human approval.

Execution rollback

For supported actions we keep history and a path back.

Agent workflow

Audit, recommendation, prioritisation and follow-up steps run as a trackable process.

Feed automation

A strong layer for working with feeds, copy, images and product signals.

Social hub

Planning and content work over social channels is part of the product layer.

Read-only client zone

After release, the client sees recommendations and exports, but doesn't change accounts or campaigns.

From the testing phase to the full version

In testingAI agents, Agent Lab, recommendations, approval inbox, execution workflow with rollback, feed automation, social hub.
Partner pilotClient audit portal, audit tokens and selected read-only agents over virtual clients — testing with first partners.
Full release (roadmap)After testing wraps up: TikTok import, CRM connector, BigQuery pipeline and the full profit layer for all e-commerce data scenarios.

Want to join the AI-agents testing phase?

Help us tune the agents before the full release. Together we'll set which agents make sense to expose to clients and which workflows stay internal under approval.