AI Assistant

The AI Assistant enables org-scoped conversations that answer infrastructure questions, leverage live platform context, and run governed MCP tools with human approval for HIGH-risk actions.

ⓘ Powered by the Axio Model Platform. See Model Platform for supported providers and routing.

What it does

▣

Answer questions

Get accurate answers from the LLM using your configured AI provider.

✓

Platform context

Include a secret-masked snapshot of your tenant to ground responses.

⚒

Run tools

Execute Platform MCP tools. HIGH-risk tools require human approval.

Assistant modes

The available experience depends on your organization's configuration.

Mode

LLM Chat

MCP Tools

Behavior

Unavailable

×

×

Configure an AI provider and/or MCP tools.

Tool runner

×

✓

Run tools by name; no conversational LLM.

AI Assistant

✓

×

Chat with the assistant only.

Assistant + tools

✓

✓

Chat and tool execution.

Chat experience

✓ Conversation list, create, delete

✓ Provider & model picker (explicit model wins over gateway route rules)

✓ Optional include platform context (default on)

✓ Markdown formatted replies

✓ Prompt Library hand-off (system guardrails + user starter)

✓ Deep links: ?prompt= and ?librarySlug=

Context Engine

Gathers a live, secret-masked snapshot for grounding and analysis. All strings pass through the shared redactor (secrets, PII).

♙User, org, project/stack counts

↗Recent/failed runs, cost, drift

♢Compliance, security findings

▦Runners, repositories

▤Recent audit

♢Deliberately omitted fields are listed in redactionFlags.

Tool calling

Tools are listed via GET /ai/tools and executed via POST /ai/tools/execute.

1

Request

User asks a question or requests an action.

→

2

Execute

Assistant validates and runs the tool.

→

3

Approve (if HIGH risk)

Human approval is required.

→

4

Result

Outcome is recorded and returned.

Tool runs are recorded and visible in Activity → Tool runs.