Short answer: usually no. Opening a chat panel automatically on every page load is the interface equivalent of a shop assistant blocking the door. It gets noticed, it also gets closed, and on mobile it hides your entire page. There are situations where it is defensible, and they are narrower than most people assume.
What auto-open actually does
The pitch is straightforward: more people see the chat, more people use it. The mechanics are less flattering. An unprompted panel arrives over content the visitor was in the middle of reading, demands a dismissal action, and on a phone occupies most or all of the screen.
The visitor's first interaction with your site becomes closing something. That is a poor opening move, and the cost is not just the click. It teaches the visitor that this site interrupts, which changes how they read everything after.
There are also concrete costs. Auto-open forces the full chat interface to load for every visitor, removing the deferred-load advantage that keeps a widget cheap on page speed. And a panel that opens and takes keyboard focus interrupts screen reader users mid-sentence with no explanation, which is an accessibility problem rather than a preference.
When it is genuinely defensible
Three cases where opening automatically earns its intrusion.
A page whose only purpose is conversation. A dedicated support page, a help centre landing page, a page you drive people to from an email that says "chat with us". Here the panel is not an interruption, it is the content.
High-intent pages with heavy dependence on a question. A booking page where availability is genuinely complicated, or a quote page where the answer depends on details. Even here, delayed and conditional beats immediate and universal.
Outside business hours, on a contact page. Someone arriving at your contact page at eleven at night is about to discover that nobody is available. A panel that surfaces immediately and offers to take their details is doing them a favour, not interrupting them.
Note the shape: all three are page-specific. Site-wide auto-open is almost never right.
The middle ground, which is where most sites should live
Between "always open" and "never open unless clicked" there is a lot of useful space.
Delay. Give visitors time to orient themselves. Something that appears after a visitor has actually read a bit of the page is far less jarring than one that arrives before the hero image has painted.
Scroll depth. Trigger after the visitor has scrolled past a meaningful point, which is a signal of engagement rather than a stopwatch.
Exit intent on desktop. Cursor heading for the browser chrome suggests departure, and offering help at that moment costs nothing since they were leaving anyway. Note that exit intent does not exist on mobile, so it needs a different rule there or none.
Page-specific rules. On the pricing page, open. On the blog, never. This is usually the highest-value distinction of all, and the easiest to reason about.
A peek instead of a full open. Show a small greeting bubble next to the launcher rather than expanding the whole panel. It communicates availability, it does not cover content, and it is dismissed with one small tap. For most sites this is the best answer.
Mobile changes the maths
Everything above is more severe on a phone. A desktop auto-open covers a portion of a wide screen; a mobile auto-open covers essentially all of it. The visitor arrives from a search result and sees a chat panel instead of the page they clicked.
Whatever your desktop rule, the mobile rule should be more conservative. Many sites are best served by no auto-open on mobile at all, or by hiding the widget on mobile entirely on certain templates. Clerkzo has both open-on-load and hide-on-mobile as behaviour toggles, so the decision is a switch rather than a code change. The wider set of mobile considerations is in chat widget mobile experience.
Frequency, memory and the second visit
If you do auto-open, it must happen once. A panel that reopens on every page navigation within a session is not proactive, it is a nuisance, and it makes navigating your site feel like swatting flies.
The rules that keep it tolerable:
- Once per session, not once per page
- If the visitor closed it, do not reopen it for that session
- If the visitor closed it repeatedly across visits, stop offering
- Never reopen after a conversation has already happened
The principle is simple: dismissal is a signal, and a widget that ignores it is telling the visitor their input does not matter. That is a strange message to send from a product whose entire purpose is listening.
What the greeting has to do if it opens itself
An auto-opened panel is an interruption, so its greeting has to justify itself immediately. A generic "How can I help you today?" is not worth the interruption; it asks the visitor to do work they did not volunteer for.
If the panel is going to open itself, the greeting should offer something specific and relevant to the page: a concrete capability, a common question, a piece of information the visitor is likely to want. On a pricing page, mention that you can explain what a project actually costs. On a service page, mention availability. The full argument for greeting copy is in what your chat greeting should say.
Suggested questions matter more here too, because the visitor did not choose to open the panel and needs an immediate reason to stay in it.
Accessibility rules for auto-open
If you use it, three non-negotiables.
Do not steal focus. The panel can appear; it must not move the keyboard user's focus into it. Focus should move only when the user chooses to enter.
Announce politely, if at all. A live region announcement that interrupts a screen reader mid-sentence is worse than no announcement.
Respect reduced motion. A panel that slides and scales into view is unpleasant for people who have asked their system to reduce motion. When that preference is set, appear instantly or not at all.
These are covered in more depth in chat widget accessibility.
The judgement call
Auto-open trades goodwill for attention. Whether that trade is worth it depends on how much the visitor wanted a conversation before you started one.
On a page where the visitor is deciding whether to buy and has an obvious unanswered question, the trade is often fine. On a blog post they arrived at from a search result, it is not, because you have interrupted the thing they came for with an offer they did not want. If you cannot articulate what the visitor is likely to be confused about on this specific page, do not open the panel on it.
The default answer for most sites is: launcher visible everywhere, panel opens on click, with a small greeting peek on your two or three highest-intent pages. That is present without being pushy, which is the whole target. Setup for all of this is toggles and a live preview; see how it works.
Frequently asked questions
Does auto-opening increase conversations?
It increases panel views. Whether it increases useful conversations depends on whether the greeting is relevant to the page. Irrelevant auto-opens mostly increase dismissals.
How long should the delay be if I use one?
Long enough that the visitor has oriented themselves and read something. Scroll depth is usually a better trigger than time, because it reflects engagement rather than the clock.
Should it auto-open on mobile?
Rarely. The panel covers the whole screen, so the interruption is total. Be more conservative on mobile than on desktop, or skip it there.
Should it reopen if the visitor closes it?
No. Dismissal is a clear signal. Respect it for the session at minimum.
Is a small greeting bubble better than opening the whole panel?
For most sites, yes. It signals availability, does not cover content, and is dismissed with one tap.