Proactive Live Chat Software: The Only Honest Number Is 2 Percent

Proactive live chat is the small popup that asks "Need a hand?" after you have been on a pricing page for a few seconds. The pitch from every vendor is that it lifts revenue. The pitch is mostly recycled.

Here is the honest version. One credible, peer-reviewed figure survives in this category: roughly 2 percent of visitors respond to a proactive invitation. The rest of the headlines vendors quote, the 305 percent ROI, the 2.8x conversion claim, the "happiness increases", are vendor studies without independent controls. If you build a model on those, you will oversize your seat count and still miss the point. The point is what your own lift is, worked out from your own traffic, not borrowed from a deck.

The 2 percent number, and why it sets the ceiling

Proactive chat is a permission interrupt. You are asking someone, who came to your site to read a page, to switch tasks and type a question. Most will not. The accepted response rate, drawn from the academic and analyst work that does exist, sits at roughly 2 percent of invited visitors.

That number sets the ceiling on everything downstream. The "x percent lift in conversion" depends on how many visitors you invite, the response rate, the value of the conversion and how your sales team handles the hand-off. None of those inputs change because a vendor published a new case study.

Run that with your own numbers. A 2 percent response rate is the working assumption, not the optimistic one. If you run a campaign and get 5 percent, you have done better than average and you should write down what you changed. If you get 0.5 percent, the campaign was poorly targeted or timed, not evidence that proactive chat is broken.

The other trap is the "lift" calculation. If you invite 1,000 people and 20 reply, you have a 2 percent response rate. If you would have made 2 of those sales anyway, and you make 3 with the chat, that is a 50 percent lift on 2 sales. It is also one extra sale. Lift percentages without a denominator are how this industry sells software.

What the meters punish

Live chat software prices by one of three meters: agent seats, conversations or tickets, and sometimes by AI resolution on top. Each punishes a different behaviour.

• Per agent. You pay for every person with a login. If you have a small team, this is the cheapest meter. It also means you pay full price during quiet weeks.
• Per conversation. You pay per chat that actually happens. A campaign that works well costs more. Tidio and Gorgias price this way.
• Per ticket. You pay per item resolved. Gorgias uses a ticket meter and adds a separate AI resolution charge that double-counts, because each AI resolution also bills as a ticket.

The meter matters more than the headline price. A $24 Tidio Starter plan looks cheaper than a $39 Asyntai Starter plan, until you count conversations. Asyntai's Starter is $39 for 800 messages. Tidio's Starter is $24.17 for 100 conversations. If your proactive campaign actually works, you blow through 100 conversations in a slow week and start paying overages or losing chats.

The published prices, side by side

These are the first-party list prices from each vendor's pricing page, September 2026. Conversions to monthly are shown for annual plans.

We make Asyntai, so read that row as a disclosure. Prices are first-party list prices from each vendor's pricing page on 2026-09-04 unless flagged.

Gorgias's prices come from published 2026 pricing analyses rather than its public page, because Gorgias publishes entry tier pricing but routes mid and top self-serve tiers through sales. The AI double-charge is the part most buyers miss: every automated resolution is both an AI resolution fee and a billable ticket.

The AI resolution fee, line by line

Three of the seven tools now meter AI responses separately. The fee is small per resolution, and adds up fast if your proactive campaign is doing its job.

The double-charge in Gorgias is the structural problem. If an AI reply resolves a customer query, Gorgias bills the AI resolution fee and also counts that conversation against the ticket allowance. One automated answer, two charges.

How the meters react when proactive chat works

The interesting case is a successful campaign. Run the 50,000-visitor example above with a 2 percent response rate: 1,000 chats, of which maybe 800 are answered entirely by an AI agent before any human sees them. Here is what each tool charges for that month, using list prices.

We make Asyntai, so read that row as a disclosure. The "800 AI resolutions" figure is illustrative. Run the maths with your own response rate.

The seat-metered tools stay cheap as long as your team does not grow. The conversation-metered tools stay cheap as long as your campaign does not work. The AI-resolution-metered tools penalise every successful automation. None of this is a problem if your campaign is small. It is a real problem if you turn the campaign on and it works, which is the case you wanted.

Which tools suit proactive campaigns at all

Proactive chat is a small, technical feature inside a larger product. Most live chat tools support it. The difference between them is whether the meter punishes the campaign that runs, and whether the AI answer is metered separately or absorbed into the plan.

For a small team that wants to test proactive chat without committing to a sales conversation, the seat-metered tools are the safest. LiveChat's Starter is $19 per agent on annual terms, $25 monthly. Olark's live-chat-only plan is $29 per agent. Help Scout's Standard is $25. The meter stays flat whether your campaign produces ten chats or ten thousand.

For a team that already knows it will run proactive chat as a permanent channel, the question is whether you want to budget per chat or per message. Tidio's conversation meter is honest about its purpose: it charges per chat, so a quiet week is cheap and a busy week is expensive. Asyntai's message meter counts every exchange; 800 messages covers roughly 800 short answers. The arithmetic above shows what each does at scale.

For a team that wants an AI agent to handle proactive chats without paying a per-resolution fee, the field is narrower. Intercom charges $0.99 per resolution on top of seats. Help Scout charges $0.75 per resolution. Gorgias charges per resolution and also counts the resolution as a ticket. Asyntai includes AI in the plan.

What proactive chat actually moves

The honest answer is "not as much as the case studies say, and more than the cynics say." The 2 percent response rate is real. The downstream lift depends on what you do with the chat once it begins. A chat that reaches a trained human who can answer in under a minute converts more often than a chat that reaches an AI that gives a useful answer, which converts more often than a chat that sits in a queue for twenty minutes.

The realistic upper bound, drawn from the public data, is that proactive chat can move conversion on the page it interrupts by a measurable but modest amount. A working assumption is a single-digit-percentage increase on the pages it runs on, with most of the lift concentrated on the small share of visitors who would otherwise have bounced. Do not budget for 305 percent ROI. Budget for the 2 percent response rate, multiply by your own close rate, multiply by your own order value, and accept that some campaigns underperform the model.

The questions to ask a vendor

1. What is your meter? Seat, conversation, message or ticket. Each punishes a different behaviour.
2. What is the AI resolution fee? If there is one, ask whether automated resolutions count against another cap as well.
3. What counts as a conversation? Most vendors count an opened chat, not a message. A 30-message chat is one conversation.
4. Can proactive invitations be capped? A bad campaign burns through your meter without lifting revenue. A daily or monthly cap protects you.
5. What happens to chats after hours? If your proactive campaign runs 24/7, your AI needs to run 24/7 as well, or you are paying humans to sleep.

What we would do

If you want to test proactive chat, start with a small monthly allowance, measure the response rate against the 2 percent baseline, and only scale once you have your own number. If your campaign works, the meter you chose starts to matter. The cheapest meter at zero conversations is rarely the cheapest meter at a thousand.

For a small business that wants proactive chat included without a per-resolution fee, Asyntai's Starter at $39 a month covers 800 messages and includes the AI agent. The conversation-metered alternatives look cheaper at the entry tier and become expensive at the scale where proactive chat earns its keep.

Asyntai Starter is $39 a month for 800 messages. Proactive invitations, AI answers and human hand-off included. No per-resolution fee. Start with the free plan to test your own 2 percent baseline before you spend anything.

Sources