ACEIRT™ Fusion
Ask AIAsk AI

Ask AI

Use Ask AI in AI Studio to move from natural-language prompts to working AI experiences, and understand when to switch to Agent Studio.

What Ask AI does

Ask AI is the conversational surface of AI Studio where you describe what you want in plain language and receive working drafts for AI-powered experiences in ACEIRT™ Fusion.

Ask AI is optimized for fast iteration: you type a prompt, see what the AI proposes, refine with follow-up messages, and then turn the result into an asset you can manage alongside the rest of your workspace.

Prerequisites

  • Access to ACEIRT™ Fusion with permission to use AI Studio
  • A project or workspace where you can save and test generated assets
  • Any domain knowledge, content, or examples you want Ask AI to base its responses on

Privacy and safety

Avoid sharing sensitive personal data, secrets, or production credentials in prompts. Treat anything you type into Ask AI as content that may be processed, logged, and used to improve the experience according to your organization and ACEIRT™ Fusion policies.

When to use Ask AI vs Agent Studio

Use Ask AI when you want to explore ideas quickly from natural language and you are not yet committed to a detailed flow design.

  • Ask AI is best for:

    • Brainstorming flows, page copy, and AI behaviors
    • Generating first drafts of agents, scripts, or widgets
    • Trying variations of tone, constraints, or instructions through follow-up prompts
    • Getting a working baseline you can later refine in other AI Studio tools

  • Agent Studio is best for:

    • Building structured, node-based flows with explicit branching and tools
    • Managing complex logic, integrations, and multi-step conversations
    • Versioning and debugging specific steps in an agentic pipeline

A common pattern is to start with Ask AI to shape the overall behavior and language, then move into Agent Studio when you need precise control over how the agent reasons, branches, and calls tools.

Get a first useful result with Ask AI

Use this quick flow to go from an idea to a result you can validate and reuse.

Frame your outcome

Start by telling Ask AI what you want to achieve, not just what to generate.

  • Describe the user, channel, and success metric: for example, “Create a voice agent that qualifies inbound sales calls and captures lead details.”
  • Mention any constraints that matter: tone, length, guardrails, or systems it must respect.

You should see Ask AI respond with an initial proposal that reflects the outcome, audience, and constraints you described.

Provide context and examples

Add the context that matters most for accurate behavior.

  • Paste representative snippets of your existing copy, scripts, or FAQs
  • Explain how your current process works and what should change
  • Call out edge cases or failures you want the AI to avoid

A stronger context lets Ask AI propose flows, messages, or behaviors that align with how your teams already operate.

Iterate with targeted refinements

Use short, focused follow-up prompts to tighten the result instead of rewriting it from scratch.

  • Ask for alternatives: “Give me three variants with more concise opening questions.”
  • Adjust constraints: “Keep the logic the same but make the tone more formal and remove emojis.”
  • Point to issues: “The current draft over-qualifies; reduce the number of questions before routing.”

After a few iterations, you should reach a draft that feels ready to test in a limited context (for example, a small traffic slice or internal review).

Decide where to take it next

Once the result is close to what you need, choose how to operationalize it:

  • For conversational agents and tool-based flows, move the logic into Agent Studio to model nodes, tools, and branches explicitly
  • For front-end experiences (like a website widget), adapt the wording and behavior into your chosen channel or template in AI Studio
  • For voice scenarios, use the same prompts and constraints as a starting point for Voice AI agents

You now have a draft that is grounded in your real context and ready to evolve into a production-grade asset.

Prompt patterns that work

Use these prompt structures to get more reliable, implementation-ready output from Ask AI.

  • Role + goal + channel
    “Act as a support assistant for our B2B SaaS product. Design a chat flow for our website widget that reduces basic billing tickets by 30 percent.”

  • Process description + constraints
    “Rewrite our existing lead-qualification script for a voice agent. Keep questions under 8 in total, avoid jargon, and always confirm consent before capturing email or phone.”

  • Input–output examples
    “Here are 5 real customer questions and how our team answered them. Propose an AI behavior that answers similarly, including when to escalate to a human.”

  • Error and edge-case focus
    “Design how the agent responds when it does not know the answer, when the user is upset, and when the user asks to speak to a human.”

  • Data-aware behavior
    “Assume the agent can look up customer plans and recent invoices via our API. Propose how it should use that data during a billing support conversation.”

Combine these patterns in a single conversation to progressively shape the AI into something that matches both your business rules and your users’ expectations.

Use these guides to take what you built with Ask AI into specific channels or more advanced flows.

Voice AI agents

Website widget embed

Build agentic flows in Agent Studio

Prompting best practices

Was this page helpful?

Last updated today

Built with Documentation.AI