Customer engagement is more than sending messages. The entire discipline is built on foundational relationship principles: showing the right experience to the right people at the right moment. Without context, that's like hitting a moving target blind.

How are teams meant to grow their customers and make the right judgement calls without customer behavior patterns? And when they do have those patterns, knowing what to build next, who should see it, when it should show up, and whether it's even doing anything once its live is an entirely new set of problems with so much access to behavioral data.

When the work spans strategy, execution, troubleshooting, and analysis, it's easy to feel more like a rubber band than a person.

Appcues AI is built to make relieve that tension and make the work easier. We have access to over a decade of in-product expertise across all the engagement principles that matter across onboarding, adoption, and engagement. When you use Appcues AI as part of your engagement workflow, you bring together your own live product data and experiences with our proven expertise.

The catch with any capable AI tool, however, is knowing what to say to it. We'll kick this article off with how to write a good prompt (feel free to skip if you're already a prompter) then walk through four of our most popular goals that you can use Appcues AI to help you achieve. Each one is anchored by one prompt you can use today in Appcues AI, and underneath each is a library of additional one-off prompts to inspire you.

First, how to talk to Appcues AI

Talking to Appcues AI is a lot like saying hi to a friend. You can't give it a secret handshake, but it uses plain language to build. Type what you want and press Enter. Keep two tips in mind to make your prompts sharper:

1. @-mentions: Type '@' to point Appcues AI at something specific. This can be a campaign, experience, workflow, segment, or key action. Instead of describing "that onboarding flow", mention it directly and Appuces AI will work from that exact item, brief and all.
2. Attachments: Drop in a file for Appcues AI to read, like a PRD, a CSV of usage data, or launch brief to build a campaign from. You can even use a screenshot of the page you're referring to, or bring in existing copy so the drafts sound like your brand. Attachments are context, not commands. Appcues AI reads them, then works through the outcome, audience, and timing with you before it builds anything.

A few habits that get better results:

• Start with the problem, not the solution. "New users aren't completing setup" beats "build me a tooltip. Let Appcues AI diagnose and pick the right experience.
• Name your audience behaviorally. "Users who signed up in the last 7 days and haven't finished the setup wizard" tells Appcues AI where someone is in their journey, which is more useful than "new users" alone.
• Describe the moment. When should this appear? Timing and targeting matter as much as the message.
• One ask at a time. Build the welcome message, review it, then ask for the follow-up. A series of focused prompts beats one giant request.

PRO TIP
For even more tailored recommendations, give Appcues AI context on your business in Settings. That's what your organization does, who your customers are, what your activation moment is, and how you'd like responses formatted.

Example: "We're a project management tool for remote teams. Our customers are opes managers at mid-sized companies. We're subscription-based with monthly/annual plans. Activation = creating a second project. Keep responses concise with actionable next steps.

Now, the four goals you can ask Appcues AI to help you with.

1. Create a new campaign

"Create a new campaign to increase activation rates for my new customers."

This is the flagship prompt we recommend. Appcues AI can take a goal, like lifting activation among new customers, and help you work backward to the campaign that gets you there. It helps you with which experiences to build, who should see them, and when. Describe the outcome you're after and it handles the diagnosis, sequencing, and targeting.

With this prompt, start broad ("new customers"), then refine the audience and the moment as you go. The prompts below break the work into its natural parts that Appcues AI can work with: deciding what to build, defining who it's for, and drafting the experience itself.

Additional prompts to inspire you

Think through the strategy

• What’s the best way to onboard new users who signed up but haven’t completed setup?
• How should I announce a new feature to power users without interrupting their workflow?
• What Appcues patterns work best for reducing churn during a free trial?
• I want to improve activation rates — what experiences should I prioritize?
• What’s the right timing and frequency for in-app messaging during onboarding?
• How can I re-engage users who haven’t logged in for 30 days?
• What’s the best approach for collecting NPS feedback?
• Should I use a modal, slideout, or tooltip to promote our new integration?
• How do I create an onboarding flow that adapts to different user roles?
• What are some proven patterns for driving feature adoption after launch?

Define and reach the right audience

• I want to target users who compelted onboarding but haven't achieved @Goal yet.
• How do I create a segment for trial users in the last 3 days?
• Build targeting rules for enterprise customers who haven't seen our new dashboard feature.
• I need to reach users who started @Flow but didn't finish it.
• Identify properties I have so I can create a segment for power users.
• How do I target users based on their subscription tier and usage?
• Build a segment comprised of users who haven't dismissed a Flow before.
• How do I exclude users who've already completed this action?

Draft the experience

• Create a Flow I can start from for new mobile app users that welcomes them to the app.
• Help me write copy for an announcement that explains our new XYZ feature.
• Give me copy for an NPS detractor follow-up email. My company's brand voice is professional and succinct.

2. Improve what's already live

"Based on how my most recently published campaign is performing against my goal, what additional tactics would you suggest implementing?"

Launching is the start, not the finish line. Appcues AI can look at how a live campaign is tracking against your goal and recommend the next move, like a follow-up experience, a sharper piece of copy, or a tactic to fill the gap. You can talk through a recommendation and apply it without leaving the conversation.

Use these when you have something live and want to make it work harder.

Additional prompts to inspire you

• What changes would make this Flow more effective at driving adoption?
• Look at this Flow and recommend ways to reduce drop-off between steps.
• Tell me about the content in this Flow. What's working and what's not?
• Help me rewrite the copy in this upgrade Flow to be shorter and more compelling.
• Review this Flow from a first-time user's perspective and call out any moments that feel like a leap.
• Suggest a simpler version of this Flow that still gets users to the same outcome.
• Help me tighten copy in this onboarding Flow without changing the meaning or the tone.
• Compare the performance of my two feature announcement variants.
• Which experiences have the highest dismiss rates and might need improvement?

3. Understand why something's happening

"Why are operators dropping off after their first week?"

When users behave in a way you didn't expect (or an experience isn't showing up the way you built it), Appcues AI helps you find the reason. Ask a plain-language question about your data and it digs into the behavior. Point it at a Flow and it explains. the exact conditions that have to be true for someone to see it, so you can catch the targeting gap or delivery bug fast.

Additional prompts to inspire you

Troubleshoot delivery and targeting

• Before I publish, can you help me verify my targeting rules will reach the right users?
• Why is my tooltip pointing to the wrong element on the page?
• Users are seeing the same announcement multiple times. How do I fix this?
• My Flow isn’t appearing even though I published it. What should I check?
• The Flow works in preview but not in production — what could be different?
• List all the experiences that may be triggering this @Flow.
• Review this flow and tell me what conditions must be true for it to appear to an end user.
• Help me identify the correct CSS selector for this element so I can target it confidently.
• I want to sanity-check that my expectations for this flow match how it’s actually configured to behave.
• Help me confirm whether this experience would show for a brand-new user vs. a returning one.

Sanity-check your audience

• Can you export a list of who’s included in XYZ Segment so I can spot check it?
• Can you explain, in plain language, who this audience actually includes today?

Find the drop-off

• Where are users dropping off in my signup-to-activation Flow?

4. Build a dashboard to see what's working

"Build me a user behavior dashboard that shows me how many active users I have, what their top behaviors are, and where they are spending time in my app."

You don’t need to jump to a separate analytics tool to understand your users. Appcues AI can build custom dashboards, like charts and funnels, from a plain description of what you want to track. Ask for active users, top behaviors, completion rates, or NPS trends, and it assembles the view so you can watch it over time.

Additional prompts to inspire you

• Which Flow with the tag ‘Onboarding’ has the best completion rate this month?
• Summarize the results of my XYZ survey Flow and highlight the key themes for me.
• Which user segment has the lowest engagement with in-app messages?
• Show me the trends in NPS scores over the last quarter.
• What percentage of users who saw the upgrade prompt actually converted?
• How long does it take users to complete onboarding on average?
• Which segment who has seen this @Flow has the highest completion rate?

What’s Appcues AI?

Appcues AI is designed to meet you where you are — whether you’re planning your next experience, improving something that’s live, figuring out why users are behaving a certain way, or building the dashboard to watch it all. These four prompts are just the anchors; the library beneath each one is there to spark your own.