WhatsApp

From Paying by the Hour to Paying for Outcomes: The Evolution of BPO with Artificial Intelligence

AI & Automation

Tabla de contenidos

BPO with artificial intelligence changes more than the technology behind the operation. It also changes what gets billed. In this blog, we walk through three stages so you can identify where your company stands today and which pricing model applies to it.

What outcome-based pricing is and how traditional BPO is evolving

Outcome-based pricing is a billing model in which the BPO, with artificial intelligence, bills for a verifiable unit of value, such as a closed sale, a resolved inquiry, or a recovered cart, rather than for hours, assigned agents, or licenses.

This shift happens in three stages:

  • In the first, traditional BPO sells capacity: you pay by the hour, per assigned agent (FTE), or per license, and cost grows in line with volume. 
  • In the second, the provider adds automation, but keeps the per-agent base and adds a variable component tied to metrics such as service level or satisfaction. Everest Group observes that most outcome-oriented agreements are still of this hybrid type, with a base rate plus bonuses and penalties. 
  • In the third, AI runs the entire process and the client pays for the outcome. This stage corresponds to the model we explain in What an AINS is and how it differs from BPO, without repeating its definition here.

To place yourself, three questions are enough: which unit appears on the invoice, who bears the risk when the service does not convert, and what happens to cost when volume goes up. If the unit is hours, the risk falls on the client, and the invoice rises with volume, then the operation is still in the first stage.

Outcome-based pricing: impact on sales and operational efficiency

Under an hourly contract, the provider bills more the more time the operation consumes; under an outcome-based one, it bills when the customer buys or the case is closed. That changes the incentive: with hours, being more efficient reduces the provider’s revenue; with outcomes, efficiency becomes part of its margin.

The market is moving in that direction, but gradually. An HFS Research survey from August 2025 of more than 500 executives anticipates that, within three years, FTE-based models will go from 42% to 28%, while outcome-based and revenue share models will rise from 20% to 39%. In software, the McKinsey analysis of 150 providers shows that outcome-based pricing is still a minority practice: only 2% of established companies use it, compared with 10% of AI-native ones.

In that same work, McKinsey clarifies when the model works: the outcome must be closely tied to the product’s work, take little time to materialize, and be confirmable automatically. A closed sale or a recovered cart meets those conditions; a diffuse process that spans several departments does not. That is why WhatsApp sales automation is a natural fit for BPO with artificial intelligence: every attempt and every close is recorded in the conversation.

AI First in conversational experience: AI first, people when it matters

Being AI First in a BPO with artificial intelligence does not mean operating without people. It means the AI agent handles standard volume by default and a person steps in for a complex objection or a sensitive complaint. Gartner is along the same lines: in a March 2025 survey of 163 service leaders, 95% plan to keep human agents, and the firm forecasts that, by 2027, half of the organizations that expected to significantly reduce their headcount will abandon that plan.

This answers the most frequent objection to a BPO with AI: “I lose control.” Control is designed through rules for handing off to an agent, contact and chat tags to analyze every conversation, and daily management reports. Chat Center operates this way in hybrid models: AI takes the volume and the human team resolves what requires judgment.

Cases and market metrics: what results each model supports

Chat Center, a Meta Business Partner, specializes in End to End automation over WhatsApp, integrates eCommerce with AI, and has managed more than 10 million chats. Its business model is based on outcomes. According to the company, its sales agents reach an average conversion rate of 18% and abandoned cart recovery reaches 25%.

In practice, BPO with artificial intelligence is measured in revenue. Movistar México, with a hybrid model of AI and people, grew its revenue 120% year over year, cut first response time to 54 seconds, and reached a 17% conversion rate. Autocompara, Santander’s insurance line, doubled its sales capacity by combining AI and human service. These are results reported by Chat Center, not market averages.

Billing should also follow the process. For sales and recovery, the natural fit is a percentage of incremental sales or a commission per closed sale; for lead qualification, a cost per qualified lead; for support, payment per resolved ticket. The details are in the article on the Revenue Share model.

How to start being AI First without dismantling your current operation

Tomorrow, review your provider’s latest invoice and identify which part corresponds to hours or assigned agents and which part to conversations that ended in a sale or resolution.

Then choose a process with a discrete outcome, such as cart recovery or lead qualification. Before talking about price, define what counts as an outcome, within what timeframe it is attributed, and which system is the source of truth. It is no minor detail: McKinsey reports that an AI prospecting company offered to charge per qualified lead or per contact made, and 90% of its customers chose to pay for activity; forecasting spend and agreeing on what a “qualified lead” is changed with each customer and made the negotiation cumbersome. Everest Group adds another risk: when provider, client, and technology partners influence the same outcome, it is advisable to agree in advance on how it is attributed.

Finally, test before migrating everything. Chat Center’s process moves through discovery, proof of concept, iteration, implementation, and expansion, with eCommerce integration so the agent prepares the cart and sends the payment link.

Find out how to turn WhatsApp into your main sales channel with End to End automation powered by AI. Book a call with Chat Center and let’s review which stage your operation is in and which pricing model fits you.

Comparte este artículo

Categorías

Suscríbete al Newsletter
Suscríbete al Newsletter

Las compañías líderes ya operan con IA

Diseñemos juntos tu estrategia.

También te podrían interesar las siguientes notas

Completa tus datos para descargar la guía