Make AI Agent improvements, new AI models, SSO login routing, and MCP tools for audit logs
Here's a quick look at the latest product improvements and fixes.
Improve Make AI Agent responses with Maia
You can now improve underwhelming agent responses with help from Maia. Review a scenariorun together, ask her what to improve, and decide whether to apply her suggested version.
Key benefits
- Improve from a scenario run: Start with the response your agent produced to focus on a specific result.
- Run-aware guidance: Maia reviews the run, the tools used, and the response to identify where changes could help.
- Targeted questions: Tell Maia what to prioritize, such as adding a missing step or improving the final response.
- Versioned improvements: Review the proposed version and choose whether to apply it.
Availability
Available on all plans.
Learn more
To learn more about how Maia helps, see Introduction to Maia by Make.
Make AI Agent: fallback connections and prompt caching
You can now set a fallback AI provider connection to use when the primary AI provider or model is unavailable. You can also enable prompt caching so the agent can reuse instructions and tool definitions, reducing latency and credit usage.
Key benefits
- Alternate AI provider connection: Set a fallback connection for your agent when its primary connection fails or is unavailable.
- Automatic retry: Make retries the run once using the fallback connection.
- Default AI provider connection: Make's AI Provider is the default fallback connection.
- Anthropic Claude model caching: Select a model caching duration when using a supported Anthropic Claude model. When you provide a conversation ID, cached context applies across repeated calls in the same conversation.
- OpenAI model caching: Supported OpenAI models automatically reuse prompt content.
Availability
Both are available in the Make AI Agent app on all plans.
Learn more
For information on all Make AI Agent module settings, see Make AI Agent (New) app.
To check out what the community is saying, see:
Claude Sonnet 5.5 is now available
Claude Sonnet 5.5 is now available in Make. Use it to process large text collections and prepare detailed drafts for scenarios that require human approval.
Key benefits
- Text input and output: Send written requests, notes, and source material to Claude Sonnet 5.5, then use its text response to classify, enrich, review, or draft content.
- Large context window: Review documents within a context window of up to 1,000,000 tokens.
- Long-form generation: Generate text outputs of up to 128,000 tokens for briefings, draft materials, and editorial review.
Availability
Claude Sonnet 5.5 is available in the Make AI Toolkit, Anthropic Claude, and Make AI Agent (New) apps.
Auto-route social logins to SSO
When an organization has a verified domain claim and SSO configured, Make routes social sign-ins for that domain into the organization's SSO flow. Users can start with Google, Apple, GitHub, or Facebook and arrive with their email already filled in.
Key benefits
- Social login routing: Users signing in with Google, Apple, GitHub, or Facebook using an email address on a verified domain are directed to the organization's configured SSO flow.
- Pre-filled email: Make carries the user's email address into the SSO flow, so users do not need to enter it again.
- Familiar sign-in choices: Users can begin with the social login option they normally use instead of finding a separate SSO option or entering an organization namespace.
- IT access control: IT admins can keep sign-ins for verified domains within the organization's configured SSO flow.
Availability
Available on all plans.
GPT-6.1 Sol is now available
GPT-6.1 Sol is now available in Make. Use it to build workflows for complex coding and professional work that gather information, prepare drafts, and route outputs for review.
Key benefits
- Tool support: Use Functions, Web search, File search, and Computer use to work with configured tools, public information, documents, and validated interface tasks.
- Complex coding and professional work: Build workflows that enrich technical requests, prepare research briefs, draft responses based on your documents, and follow up on defined operations tasks.
- Multi-agent support in beta: Coordinate defined subtasks and consolidate outputs into an implementation plan draft for expert approval.
- Clearer boundaries: GPT-6.1 Sol is designed to respect your intent and safety constraints more reliably, and to be more transparent about its limitations.
Availability
GPT-6.1 Sol is available in the Make AI Toolkit, OpenAI (ChatGPT, Whisper), and Make AI Agent (New) apps on all plans.
Learn more
To check out what the community is saying, see App update: Automate processes with GPT-6.1 Sol in Make.
MCP tools for audit logs are now available
You can now query audit log events through MCP. With MCP tools for audit logs, organization and IT admins can bring answers about scenario changes and organization activity into the agents and apps they already use.
Key benefits
- List audit log events: Retrieve audit log events through MCP.
- Filter audit events: Narrow results by user, action, or date.
- Inspect individual entries: Retrieve the details of a specific audit log event.
- Read-only access: Query audit information without making changes to your organization.
Availability
Available on all plans.
Learn more
To learn how to view the MCP tools available through Make, see Make MCP Server.
To check out what the community is saying, see Feature Spotlight: MCP tools for audit logs in Make.