Trust and control
Your rules. Your data. Your playbook.
May was about giving teams the confidence to go deeper with Amp: knowing the platform works the way they expect, handles data the way compliance requires, and presents itself with the polish users deserve. From a reimagined chat workspace to enterprise-grade data governance, this release is built around trust: in the product, in the process, and in the AI doing the work.
The headlines 06
Modern chat workspace
A workspace that feels as intelligent as the AI inside it.
Playbooks, built for teams
Save, name, and switch between hiring playbooks instantly.
Configurable data retention
GDPR, CCPA, SOC 2: your compliance rules, enforced automatically.
Integrations hub
Every integration, in a single place.
MCP server for Claude.ai
Bring Amp's recruiting intelligence to Claude.ai custom connectors.
Bias testing per job
See fairness issues before they become hiring problems.
Notable improvements 05
Plan mode for workflow analysis
Agents can now enter Plan Mode to analyze a workflow, generate a structured proposal, and let you review it before any action is taken. A clear checkpoint between intent and execution, especially valuable for complex, multi-step hiring workflows.
Collapsible tool call indicators
When an agent is working through a task, tool calls now appear as a collapsible indicator with plain-English action names instead of raw function labels. The workspace feels cleaner, and it's easier to follow what the agent is actually doing.
Archived agents in the workspace
Archived agents now appear in the workspace dropdown with a clear restore action, so nothing is ever truly lost. Teams can archive unused agents knowing they can bring them back with a single click.
ATS stage auto-update on interview confirmation
When an interview is confirmed, Amp now automatically advances the candidate to the correct ATS stage, removing a manual step that was easy to forget and keeping your pipeline data accurate.
Secure file uploads with virus scanning
All file uploads now go through magic-byte type validation and automated virus scanning before they're processed or surfaced to agents. Files are blocked while a scan is pending, so your workspace only ever works with clean, verified documents.