watsonx Orchestrate

Hi, thank you for your question; can you first clarify why you are using our legacy assistant tech and why you could not do it all with an agent?

What you tried (adding the Assistant as a collaborator inside the Agent) won’t work, because in watsonx Orchestrate the AI Assistant and the AI Agent are two different building blocks.

• AI Assistants are built for conversational flows (NLU + intents + dialog).

• AI Agents in Orchestrate are designed to trigger and orchestrate “skills” (actions, workflows, APIs).

Right now, there isn’t a direct way to “plug” an Assistant into an Agent so that the Agent just reuses all of the Assistant’s prompts/dialog. They don’t automatically share context.

✅ What you can do:

1. If you want a chatbot with both dialog + orchestration, you need to expose your AI Assistant as an endpoint (Skill) and then connect that into your Agent.

2. Alternatively, rebuild the Assistant’s prompts/intents inside the Agent as Agent skills, so the Agent can handle them natively.

3. IBM’s roadmap is to make this integration smoother, but today it requires either API linking or manual recreation of dialog logic.

So, the reason it didn’t work is because collaboration does not mean inheritance of dialog. To connect them, you’ll need to either wrap the Assistant as a callable skill, or redesign the flow inside the Agent.

To connect an AI Agent with an AI Assistant, you can:

1. Use APIs – Set up communication between both using REST or WebSocket APIs.

2. Middleware – Use a backend (e.g. Node.js, Python) to handle logic between them.

3. Same Platform – If both are on the same AI platform (like OpenAI, Google, etc.), check for built-in integration options.

4. Webhook Triggers – Set one to trigger actions in the other based on specific events.

Let me know what tools or platforms you’re using — I can give more specific steps.