The Intelligent Assistant is the chatbot embedded in the AAP UI, and as of AAP 2.7 it supports a BYOM (Bring Your Own Model) provider matrix — Red Hat AI, OpenAI, and Azure OpenAI can all serve as its backend. This example wires the Intelligent Assistant to Red Hat AI using the AAP Gateway configuration API.
Example Playbook
---
- name: Configure Intelligent Assistant with Red Hat AI backend
hosts: localhost
gather_facts: false
vars:
aap_gateway_url: "https://aap.example.com"
aap_gateway_token: "{{ lookup('env', 'AAP_GATEWAY_TOKEN') }}"
rhaiis_endpoint: "https://rhaiis.internal.example.com/v1"
rhaiis_model_name: "granite-3.1-8b-instruct"
tasks:
- name: Register Red Hat AI as an assistant model provider
ansible.builtin.uri:
url: "{{ aap_gateway_url }}/api/gateway/v1/ai_providers/"
method: POST
headers:
Authorization: "Bearer {{ aap_gateway_token }}"
body_format: json
body:
name: "redhat-ai-primary"
provider_type: "red_hat_ai"
endpoint_url: "{{ rhaiis_endpoint }}"
model: "{{ rhaiis_model_name }}"
verify_ssl: true
status_code: [200, 201]
register: provider_result
- name: Assign the provider to the Intelligent Assistant
ansible.builtin.uri:
url: "{{ aap_gateway_url }}/api/gateway/v1/settings/intelligent_assistant/"
method: PATCH
headers:
Authorization: "Bearer {{ aap_gateway_token }}"
body_format: json
body:
ai_provider_id: "{{ provider_result.json.id }}"
enabled: true
status_code: 200
- name: Confirm the assistant is reachable
ansible.builtin.uri:
url: "{{ aap_gateway_url }}/api/gateway/v1/settings/intelligent_assistant/status/"
method: GET
headers:
Authorization: "Bearer {{ aap_gateway_token }}"
status_code: 200
register: status_check
- name: Show assistant status
ansible.builtin.debug:
var: status_check.json
What This Does
The play calls the AAP Gateway API to register a Red Hat AI endpoint as an ai_providers entry, then patches the intelligent_assistant setting to point at that provider ID and enable it. A final GET confirms the assistant can reach the model. Red Hat AI is supported as an Intelligent Assistant backend from AAP 2.6 onward, alongside OpenAI and Azure OpenAI — this was reaffirmed in Red Hat's BYOM provider compatibility matrix presented at Red Hat Tech Day Netherlands 2026 in Bunnik. Note that this is a different matrix than the one for the Coding Assistant (the Ansible VS Code extension), which has its own provider support timeline.
Notes / Gotchas
- IBM watsonx is not supported as an Intelligent Assistant backend at all — it only shows up on the Coding Assistant side of the matrix, where it was actually the first external provider supported (AAP 2.5+). Don't assume parity between the two assistants' provider lists.
- Google Gemini/Vertex is "Coming Soon" for the Intelligent Assistant as of this writing — plan a fallback provider if your organization standardizes on Gemini.
- Store
aap_gateway_tokenin Ansible Vault or your CI secret store — never commit it inline. provider_typeand endpoint field names are illustrative; confirm exact Gateway API field names against your AAP 2.7 instance's API schema (/api/gateway/v1/) before running in production.- If you also plan to use BYOK (Bring Your Own Knowledge, currently Tech Preview), that's a separate configuration step — it feeds the Intelligent Assistant's RAG pipeline with your organization's own change-management procedures, naming conventions, and compliance runbooks (PCI-DSS, HIPAA, SOC2), rather than changing which model answers the questions.
Once the provider is registered and assigned, the Intelligent Assistant routes every chat request through your Red Hat AI endpoint — verify connectivity with a status check before rolling this out to a full AAP environment.
Related Articles
- Ansible BYOM Providers Compared: Choosing a Backend for AAP AI Features
- Ansible BYOK Example: Add Company Runbooks to the Intelligent Assistant RAG
- Ansible Intelligent Assistant Example: Configure OpenAI as the Backend