Skip to main content
Before configuring an assistant, decide what experience you want visitors to have. A clear purpose helps you choose the right knowledge, conversation style, and next step. An AI Experience is more than a place to ask questions. It gives people an opportunity to interact with your expertise, understand their options, and decide whether they want to continue the conversation with you.

1. Start with the visitor’s problem

Describe the situation that brings someone to your website. Focus on what they need to understand or decide, rather than on the technology you plan to use. Ask yourself:
  • What questions do prospects repeatedly ask before contacting me?
  • What makes it difficult for them to evaluate their options?
  • Which part of my expertise would be most useful before an initial meeting?
Example: A business consultant notices that prospects often struggle to identify which operational problems are worth solving first. Their AI Experience could help visitors clarify the problem and explore possible approaches—not promise a complete consulting engagement in chat.

2. Identify your intended audience

Choose the people the experience is designed to help. You can serve multiple audiences, but begin with one primary group so the conversation remains focused. You do not need a detailed demographic profile. The visitor’s situation, questions, and goals are often more useful than a broad industry label. For a deeper walkthrough, see Describe your audience.

3. Define the value visitors should receive

Complete this sentence:
My AI Experience helps [audience] understand or explore [problem or decision] using [my expertise or approach], so they can [useful outcome].
For example:
My AI Experience helps business owners explore operational bottlenecks using my consulting framework, so they can better understand where to focus before booking a discovery call.
A useful outcome could be:
  • Understanding unfamiliar concepts in plain language.
  • Comparing general approaches and their trade-offs.
  • Clarifying goals, priorities, or questions.
  • Preparing for a more productive conversation with an expert.
Avoid defining success as simply “answering any question.” A focused experience is easier to explain, test, and improve.

4. Choose the role of your assistant

Decide how the assistant should help visitors. Common roles include: An experience may combine roles. For example, it can explain a topic first and then guide the visitor through a few reflective questions.
The assistant should not imply that it can perform services, access live information, or make decisions that your setup does not support.

5. Set clear boundaries

Specify what the AI can and cannot do. This protects visitor trust and helps the assistant recognize when a human should take over. For example, an insurance education experience might:
  • Help with: Explaining general insurance terminology and differences between common coverage types.
  • Not do: Collect sensitive medical details, determine eligibility, or present general information as personalized professional advice.
  • Hand off when: A visitor wants a personalized recommendation, quote, or application assistance.
Other experiences may need boundaries around live inventory, current pricing, legal or medical advice, account-specific information, or confidential data.
Do not invite visitors to submit sensitive personal information unless you have an appropriate, verified process for collecting and protecting it. A conversational AI is not automatically suitable for confidential or regulated information.

6. Decide what a successful next step looks like

An AI Experience should provide value even if a visitor never becomes a lead. When someone is ready to go further, however, make the human next step clear. Depending on your business, that might mean:
  • Booking an introductory conversation.
  • Contacting your team about a specific need.
  • Exploring a service or resource on your website.
  • Returning later with better-defined questions.
The assistant should recognize interest naturally rather than pushing every visitor to book a meeting immediately.

7. Write your one-page experience brief

Use this template before moving into configuration:

Example: An AI Experience for a leadership coach

Audience: Managers preparing for difficult team conversations. Visitor problem: They are unsure how to approach a conversation constructively. Value: Explore a coaching framework, reflect on the situation, and identify questions to prepare. Assistant role: Educational guide grounded in the coach’s published methodology. Boundaries: Does not diagnose mental health conditions, claim to know the other person’s motives, or replace a professional coaching relationship. Human next step: Invite a coaching conversation when the visitor wants personalized support.

8. Check your concept before building

Your initial concept is ready when you can answer yes to these questions:
  • I can describe the primary audience in one sentence.
  • I know what problem or decision the experience helps with.
  • I can explain the value without mentioning AI technology.
  • I have identified knowledge or expertise to support its answers.
  • I have defined important limitations and handoff situations.
  • I know what a useful visitor outcome looks like.
You can refine these choices after testing. The goal is a clear starting point, not a perfect specification.

What’s next?

Continue with Describe your audience to understand the visitor’s needs in more detail. Then use Design the conversation to turn your concept into a helpful interaction. If you’re ready to configure an assistant directly, see Create your first AI Experience.