← Back to Blog

Screen-Aware AI Onboarding Assistant: How It Works

A screen-aware AI onboarding assistant is an in-app chat that knows where a new user is in your product (the page, the active tab, any open dialog) and answers their “how do I...?” questions from that spot, so they never have to describe what they’re looking at. In Total Chat’s version, for logged-in users, it can also take them to the right page and point out the control they need. That’s the whole idea: onboarding help that starts from where the user already is, when they ask for it, instead of a tour you scripted months ago and hope they sit through.

Last updated: September 2026

This changes what you plan for

Most SaaS onboarding is built on a guess. You guess which five features a new user needs, script a tour in that order, and ship it. Then the UI changes, a button moves, and step three points at empty space.

A screen-aware assistant flips the order. The user asks first, and the answer is built from the screen they’re on. You stop maintaining a script and start maintaining a knowledge base and a map of your features.

I think that’s the right trade for most small teams. Frankly, a five-step welcome tour is something people click through to reach the product. The questions they type later, stuck on a real task with real data, are the moments that decide whether they stay.

What “screen-aware” means in practice

The phrase gets stretched, so here’s the concrete version. With every message a logged-in user sends, Total Chat’s assistant receives:

• the current page
• any modal that’s open
• the active tab
• up to 10 features available on that page, taken from a scan of your app’s code
• whether they’re on mobile or desktop

What it doesn’t get matters just as much. No screenshots, no copy of the full page, no form values or filter settings. It knows which page, tab and dialog someone is on and what they can do there. It does not see their screen, and I’d be wary of any vendor who implies theirs does without explaining how.

For onboarding, that context removes the most tedious line in any support exchange: “Which page are you on?” A new user asking “where do I add a teammate?” from a settings page gets a different answer than the same question typed on the dashboard, because the assistant knows the difference.

From answer to action

Knowing the page is half of it. Acting on it is the other half.

Total Chat can run a guided walkthrough of one feature. It moves the user to the right page, waits for it to load, points out the control they need, and explains the last step in chat. If the feature lives behind a button or a dialog, it clicks the control that opens it. The walkthrough survives a page reload, so moving between pages doesn’t lose the user’s place.

There are limits, and you should know them before you plan around this:

1. One feature per walkthrough. Chaining several features into one workflow (“set up billing, then invite your team, then connect your calendar”) isn’t built.
2. It won’t fill in forms. Users type their own data.
3. Pointing out a specific control depends on the code scan having captured that element’s selector. So the accurate phrasing is that it can mark the exact control, not that it always will.

For a closer look at how walkthroughs compare with scripted tours, see AI User Onboarding Walkthrough: What Actually Works.

The caveat to hear before you buy

The full version of this (page context beyond the URL, walkthroughs, and pointing out controls) runs on the feature map that Total Chat’s codebase scanner produces. That scanner isn’t available as a self-serve download yet. Without it, the assistant still knows which route a user is on and answers from your knowledge base, but it doesn’t have the per-page feature list it needs to guide.

Until the scanner is published, there’s no way to build a feature map for your app yourself. Email [email protected] and ask before you plan around it. The scanner also reads Next.js apps only, so if your product runs on another framework, walkthroughs and control-pointing aren’t available for it today.

I’d rather you read that here than find out after install. If your onboarding plan depends on walkthroughs from day one, ask about the feature map first.

A second limit: all of this is logged-in only. The pre-login widget on your marketing site answers from the knowledge base, captures leads and can hand a conversation to a person, but it has no page awareness and doesn’t run walkthroughs.

Setting one up, in order

Install the widget. Total Chat ships as a script tag with streaming responses, not an SDK you wire into your build. It goes in your app’s root layout.

Seed the knowledge base. Write articles yourself, or point the URL crawler at your live site; it drafts 3 to 8 articles for you to approve or reject. Onboarding answers are only as good as what’s in there, so read those drafts properly rather than approving them in bulk.

List your first-week questions. Pull them from support email, sales calls and churn surveys. If one isn’t answered in the knowledge base yet, write that article before launch.

Decide what happens when the assistant can’t help. Set an escalation email and it can hand the conversation to a person, who gets the full transcript while the user sees a handoff note. That’s an email handoff, not a live-chat takeover.

Watch for gaps. On paid tiers, when a conversation exposes something your knowledge base doesn’t cover, the assistant drafts an article. Nothing goes live until you approve, edit or reject it.

New users also find bugs. When someone says a feature is broken, the assistant files a structured bug report with the page address, browser and recent JavaScript errors, on every plan.

Screen-aware assistant vs. scripted product tour

	Scripted tour	Screen-aware assistant
Starts when	You decide (usually first login)	The user asks
Knows the current page	Only the page it was built for	Yes, on every message
Covers questions you didn’t predict	No	Yes, if the knowledge base answers them
What you maintain	The script	The knowledge base and feature map

Tours still have a place. For one must-do setup step, a fixed tour is fine. Intercom’s Product Tours, for example, are pre-scripted and sit on its $85/seat Advanced plan. The case for a screen-aware assistant is everything after that first step, when users go off-script.

For the support side of the same idea, read What a Screen-Aware Customer Support Chatbot Actually Does and Why Screen-Aware AI Is the Future of SaaS Customer Support. Install details are in the Total Chat docs.

Frequently Asked Questions

What is a screen-aware AI onboarding assistant?

It’s an in-app chat that knows which page, tab and open dialog a new user is on and answers their setup questions in that context. Total Chat’s version, for logged-in users, can also take the user to the right page and point out the control they need, so onboarding help starts from where they already are rather than from a pre-scripted tour.

Does Total Chat see the user’s screen?

No. With each message it receives the current page, any open modal, the active tab, the features available on that page, and whether the user is on mobile or desktop. It doesn’t take screenshots, copy the full page, or read form values or filter settings. It knows where the user is and what they can do there.

Can the assistant complete onboarding steps for the user?

Partly. It can run a walkthrough of one feature: take the user to the page, point out the control and, for a button or dialog, click the control that opens it. It explains the final step in chat. It doesn’t fill in forms, and it can’t chain several features into a single workflow yet.

Does it work on my marketing site before users log in?

Not for onboarding. The pre-login widget answers from your knowledge base, captures leads, and can hand a conversation to a person by email when you’ve set an escalation address. Page awareness, moving users between pages and walkthroughs are logged-in features only, so onboarding guidance starts once a user is inside your app.

How do I get started?

Total Chat installs as a script tag, not an SDK. A free workspace is set up as soon as you sign up: your App ID and API key are shown once and emailed to you, and they’re also your dashboard login. Walkthroughs rely on a feature map from the codebase scanner, which isn’t a self-serve download yet, so email [email protected] if you need one.

---