what is a "resolved conversation" and how is it billed?

There are three common ways to bill this kind of software: per message sent, per user/agent with access to the system, or per resolved conversation. Each aligns the vendor's incentive differently:

• Billing per message punishes long or complicated conversations — exactly the ones that need the most back-and-forth to resolve well. An undecided customer who needs five questions before buying ends up costing more than one who resolved everything in a single message, without reflecting any extra value delivered.
• Billing per user (human agent seat) has no direct relation to how many people were served — a business with 3 agents can serve 50 customers or 5,000 and pay the same. It also penalizes adding people to the team, which misaligns pricing from real growth.
• Billing per resolved conversation ties the cost to what actually matters: how many distinct customers were served. It doesn't matter whether the answer took 2 messages or 20, or how many team members took part — it counts once per customer every 24 hours.

This also aligns the vendor's incentive: there's no benefit to artificially stretching a conversation (as there would be if billing per message), because the price is fixed the moment that conversation opened with that customer.

When a customer writes for the first time in a day, a conversation opens. Everything that happens with that same customer over the next 24 hours — ten messages or a hundred, one or three human agents stepping in, the AI agent on autopilot or copilot — keeps counting as that same conversation. If the same customer writes again 30 hours later, that opens a new conversation, because the window has passed.

This criterion is conceptually the same one Meta uses for its own per-conversation billing on the WhatsApp Business Platform — it's not something invented by support platforms, but a de facto standard in the business messaging industry.

When comparing prices across platforms, it helps to translate everything to a single denominator: cost per resolved conversation per month, instead of eyeballing plan prices. Useful questions for a fair comparison:

• How many conversations does the plan include, and how much is each additional one (overage)?
• Is there a setup fee separate from the monthly plan?
• Does the price rise if you add channels or users, on top of conversation volume?
• Does the cost per extra conversation drop on larger plans, or stay flat?

With those four answers, the real expected monthly cost — base plan plus likely overage — becomes comparable across vendors, instead of comparing only the entry price.

heysteff bills on exactly this basis: a conversation is one customer served within 24 hours, on any channel. Each plan comes with a number of included conversations and a decreasing overage cost as the plan grows — from 0.5 USD per extra conversation on Starter to 0.3 USD on Max — with no setup fee on any plan, and 20% off on annual billing.

• What Is Conversational Commerce?
• WhatsApp Business Platform vs. Business App
• heysteff comparisons
• Conversational AI glossary