Supermetrics 扩展 Claude 营销工作流并加入审批控制

Supermetrics announced on November 3, 2026 an expanded Claude marketing workflow covering analysis, campaign creation and editing, and reported that more than 6,000 customers have used the connector.

The announcement names campaign creation or editing for Google Ads, Meta, Microsoft Advertising, TikTok, LinkedIn and Snapchat. Claude can draw on connected advertising, CRM, SEO, ecommerce, data-warehouse and other sources, with Supermetrics supplying normalized marketing data when campaign information is requested. New campaigns remain paused until they are reviewed and explicitly enabled, and marketers can review proposed edits before they are applied.

Supermetrics said active users have grown by an average of 250% month over month since the connector’s February 2026 launch, with paid-media reporting currently the dominant use case and the connector ranking among its fastest-adopted product launches. That makes analysis the entry point for execution: performance investigation and review of a campaign proposal can happen within one Claude conversation. Co-CEO and founder Mikael Thuneberg’s stated rationale is to reduce the operational work required after Claude identifies an opportunity, without surrendering control of campaigns or budgets.

The current campaign-management documentation labels the feature as early access. It requires an active Supermetrics trial or subscription with an AI chat destination, and campaign write access must be granted separately for each ad account. The integration does not increase a user’s permissions at the ad platform, so view-only platform access still prevents changes.

Teams can require approval before a campaign goes live, before budget changes on live campaigns and going live, or for every change to a live campaign. A separate no-human-approval setting applies changes immediately, so human review is configurable rather than universal.

Every change is logged in Campaign History. Supported previous values can be restored, including budget, status, targeting and ad settings, but removed campaigns cannot be recovered. Editing also varies by platform; Snapchat permits changes only to an ad’s name and status, not its creative.

Business Context Across AI Tools

The Context Layer carries targets, reporting rules, client-to-account mappings, naming conventions and events into AI analysis and campaign execution. Teams can define these rules in natural language, avoiding repeated explanations of their marketing setup in each conversation.

The Business Context documentation describes shared team memory built into the Supermetrics MCP server and used by Claude, ChatGPT and Supermetrics Studio. Saved guidance can cover reporting time zones, currencies and attribution models; analysis exclusions; campaign naming; brand language; client mappings; and known events such as tracking outages or budget changes. Context can apply to a whole team, a data source or an account. Only team owners, admins and editors can add, change or delete it, and account-level context follows the same access restrictions as the underlying data.

These components predate the November 3 announcement. Business Context documentation was published on September 24, 2026 and updated on October 7, 2026. A May 14, 2026 workflow post had already described starting, pausing and ending campaigns, refining targeting and adjusting budgets through Claude, with previews, paused campaign creation and action logging. The new development is therefore the expanded workflow and adoption disclosure, not the first appearance of campaign actions or saved business context.

Layer’s Reporting Workflow

A Supermetrics case study supplies an adoption example on the analysis side. Norwegian agency Layer reported that some bespoke client reports requiring up to 10 hours of manual spreadsheet work now take approximately 20 minutes with Supermetrics and Claude. Seventeen of those minutes are spent placing the prepared data into the client’s final Google Sheet. Layer continues to use Looker Studio dashboards alongside conversational analysis for custom and multi-channel requests.

Before broader client use, Layer’s Morten Kleven spent several weeks testing and validating the AI output. The case study reports that this process found no hallucinations or incorrect data points. The example concerns reporting rather than campaign execution, with a manual handoff and existing dashboards still part of the workflow.

The practical shift is a conversational path from cross-platform data to a proposed campaign change, with saved business rules carried into the exchange. The reported adoption figures establish use of the connector, while the documentation sets the conditions for action: early-access campaign tools, explicit write access, platform permissions and the approval level each customer selects.