A finance manager must select a Claude feature for a task that requires deterministic calculations across a multi-tab spreadsheet.
Which selection best fits the requirement?

• A. Use the Research feature for the calculations, which gathers and synthesizes external sources and is not the appropriate feature for deterministic computation on an uploaded spreadsheet.
• B. Use Code Execution, which runs code in a sandbox to perform calculations on the uploaded file and produce outputs the associate can verify.
• C. Use a standard chat prompt for the calculations, which generates text without running code and does not produce the reproducible computational output the multi-tab calculation task requires.
• D. Use an Artifact alone for the calculations, which provides a stable editable surface for the result but does not perform the deterministic computation that Code Execution would run.

A knowledge worker is preparing to use the Gmail Connector with Claude for the first time.
Which statement most accurately describes the Connector's behavior at the user level?

• A. Claude searches and reads emails the user has access to and labels them according to user-defined rules, with reply drafting handled in a separate workflow.
• B. Claude searches and reads emails the user has access to and drafts replies for the user to send manually, with each action requiring explicit user approval.
• C. Claude searches and reads emails the user has access to and sends replies on the user's behalf for low- risk messages, with higher-risk messages routed to the user for approval.
• D. Claude searches and reads emails the user has access to in a read-only mode, and the user must compose any reply manually outside the Connector.

An analyst is working in a long Claude conversation. The conversation contains roughly 60 turns covering an initial requirements analysis, several rounds of data review, and a current drafting phase. Recent responses have started to drift from the agreed requirements.
Which step most directly addresses the drift while preserving the analyst's progress?

• A. Open a new chat and paste only the current draft, then continue refining it there without the prior requirements and data-review context.
• B. Switch the conversation to a model with a larger context window so the accumulated history no longer causes drift, and continue without restructuring the inputs.
• C. Move the source data and requirements into a Project knowledge base, then continue the drafting work in a new chat inside that Project.
• D. Summarize the agreed requirements, key data findings, and current draft state, then continue in the same chat by reasserting the summary as the working context.

A knowledge worker is explaining the purpose of an organizational AI governance framework to a colleague.
Which description is most accurate?

• A. A governance framework lists the AI tools the organization licenses and the cost of each, so users can pick the most cost-effective tool for each task.
• B. A governance framework documents the AI policies the organization is legally required to adopt, focusing on regulatory compliance rather than day-to-day use.
• C. A governance framework defines the training every employee must complete before using AI tools, with separate requirements for each tool the organization licenses.
• D. A governance framework defines approved tools, allowed data, required reviews, and reporting paths so AI is used consistently and accountably.

You are setting up a Connector for a shared data source that several colleagues will use.
Which Connector setup step should be performed first?

• A. Confirm that the Connector is approved and that the data classification is appropriate.
• B. Use the Connector in a representative task to evaluate its real-world utility.
• C. Connect the data source and confirm that authentication is working as expected.
• D. Document the Connector's purpose and the appropriate use cases for the team.