Blog

Why We Charge Per Completed Qualification, Not Per Message

Every pricing model creates an incentive. Most AI chatbot pricing models create the wrong one.

November 2026 · 5 min read

Ask any AI chatbot vendor how they price, and you will get one of three answers. Per message, per user seat, or per qualified lead. Each one tells you something about what the vendor actually optimises for.

LeadWise AI charges per completed qualification. Not per message, not per conversation, not per seat. Here is why that matters — and what it does to the incentives on both sides.

The short version: you should only pay when the platform produces something you can act on. Everything else is bookkeeping.

In this article

The two models we rejected

Per-message pricing sounds fair until you think about what it measures. It counts how many times the chatbot said something, regardless of whether the prospect was engaged or already leaving.

The problem is that a prospect who abandons after one message and a prospect who abandons after twelve both cost you the same in real terms — nothing. But per-message pricing charges you twelve times for the second one. It penalises long conversations that never became leads, which is exactly the wrong incentive for a tool built to have long, careful conversations.

Per-seat pricing is the other common option. It measures how many people on your team can log into the dashboard, regardless of whether those people are actually doing anything with the leads. It is a pricing model borrowed from CRMs, where seat count approximates usage. For a qualification engine, it does not approximate anything useful.

What happens to abandoned conversations

Nothing. They are free.

When a prospect starts a conversation and then leaves, that conversation is recorded for analytics purposes — you can see exactly where in the flow they dropped, and which question caused it — but it is never billed. It only appears in your invoice if the prospect came back later and completed the qualification.

That means you can run a candid experiment with your qualification questions. If a question is losing 40% of prospects, you can see it, change it, and measure the difference. You are never paying to run that experiment.

Volume discounts apply where you would expect

Because the per-lead fee is the unit, volume discounts apply to the marginal leads — the ones above the first band. Not to the whole month.

The first 25 leads each month are always at full price. Leads 26 to 100 are 15% off. Leads 101 to 500 are 25% off. Leads 501 and above are 35% off. Each discount applies only to the leads inside that band.

This is the standard way pricing tiers work when a cost structure has both fixed and variable components. The first band covers the base fee. Higher bands are pure marginal cost, so they can be discounted more aggressively.

How to evaluate any pricing model

Four questions to ask any vendor whose pricing model you are considering.

  1. What does the model charge you for — messages, seats, conversations, or completed outcomes?
  2. Does the unit it charges for correspond to something your business actually values?
  3. Who bears the cost when the platform underperforms — you or the vendor?
  4. Can you predict your monthly invoice before the month starts, or do you only find out at month-end?

See what this looks like for your business

We would be glad to walk through our pricing model and what your specific cost per qualified lead would look like. No pitch — just a conversation about the numbers.

Contact Us See Pricing

← All articles

To Top