Why Latin American buyers already use AI, but don't trust your company to
In Latin America, 6 in 10 consumers already use AI daily, yet distrust companies that use it without saying so. What that changes in a sales call.
Cristian Pereyra
Co-founder · Engineering
A salesperson selling an AI agent opens the demo by explaining what a conversational agent is, how it differs from a chatbot, and why it isn't magic. The prospect, meanwhile, used ChatGPT that same morning to draft an email. The meeting burns time explaining something the other person already knows, and the time left for the real objection (who reviews what this says before it reaches one of my customers?) disappears.
The problem isn't that your prospect never used AI
A 2025 Bain & Company survey found that 65% of consumers in Latin America already use artificial intelligence, led by Colombia and Peru. In Brazil and Mexico that figure climbs to 70-76% when asked specifically about generative AI tools like ChatGPT (Omdia, Digital Consumer Insights 2024), above the global average. Assuming you need to explain what a language model is in the first meeting isn't caution. It's misreading who's sitting across from you.
Artificial intelligence use
65/ 100
- already use it
- not yet
Bain & Company, survey of Latin American consumers, 2025. The figure covers self-reported use of AI in general, not generative AI alone.
What's missing isn't knowledge, it's trust
That same Bain research describes a paradox: Latin Americans use AI, but don't fully believe in it. 44% worry it will spread false information, 38% fear losing their job to it, and 35% still don't feel ready to use it. Globally, 6 in 10 consumers say trust will be the single most important attribute of any company as AI advances (a Salesforce figure, cited in the same Bain study). That distrust isn't aimed at the technology. It's aimed at the company using it without showing how.
44%
worry AI will spread false information (Bain & Company, 2025)
38%
fear losing their job to it
35%
don't feel ready to use it
Why the prospect wants proof, not an explanation
A WhatsApp Business report with Kantar, surveying 11,056 consumers across 22 markets (including Mexico, Brazil, Argentina, and Colombia), found that 79.3% need signals of legitimacy before interacting with an automated channel. This isn't generic skepticism. It's a specific question about who's on the other end and what happens if something goes wrong. It's the same question that shows up in a sales meeting when the prospect asks "what if it says something wrong to a customer?" and gets answered with another demo feature instead of the actual control mechanism.
79.3%
need signals of legitimacy before interacting with an automated channel (WhatsApp Business with Kantar, 11,056 consumers across 22 markets)
74.6%
trust a company more if they can exchange messages with it
What changes in the meeting once you understand this
That same report found something directly useful for sales: 74.6% of consumers trust a company more if they can exchange messages with it. WhatsApp is already, for most prospects in the region, the channel where that trust already lives; penetration in Mexico is above 90% (Blip, 2026). Selling an agent that runs there, on the channel a prospect already uses to talk to their bank or their doctor, starts with less resistance than selling a brand new chat widget on a website.
The second thing that changes is the order of the demo. In more than one meeting, the moment the prospect relaxes isn't when the agent answers a hard question well. It's when they get shown the screen where every conversation is logged, and who can step in if something goes off script. Showing the brake before the accelerator answers the question they actually had, even if they never said it out loud.
The same 45-minute meeting, in two orders
Explanation first
The objection lands when there's no time left.
- 0-15What an agent is and how it differs from a chatbot
- 15-35Demo: the agent handles hard questions well
- 40"What if it says something wrong to a customer?"
- 45Answered with another demo feature
Control first
The objection is already answered when it arrives.
- 0-10Where every conversation is logged and who can step in
- 10-15The objection shows up here, already answered
- 15-35Demo: the agent handles hard questions well
- 35-45Time left for the prospect's own case
The agenda is illustrative. What isn't: the objection almost never gets said out loud, so the order decides whether there is any time left to answer it.
The sales conversation that actually works
Selling conversational AI in Latin America doesn't require convincing anyone the technology works. Most people already use it every day, even if only to write an email. It requires showing, before anyone asks, who controls what the agent says when nobody's watching. That shift in focus, more than any new demo feature, is what moves the meeting forward.
The first ten minutes, concretely
Three screens, in this order: the conversation log, with a real one open; who on their team can step in, and when; and what the agent does when it doesn't know something, showing an actual handoff rather than describing one. The demo comes after that. If the prospect interrupts with the control question first, you're late by a minute instead of by half a meeting.
If you're curious how that control gets built on the inside, not just how it gets shown in a sales meeting, there's more on the topic in Why telling an AI agent "don't do that" is not a security strategy.
Sources:
- Bain & Company: How are Latin American consumers thinking, feeling and acting? (2025)
- Omdia, Digital Consumer Insights 2024 (generative AI use in Brazil and Mexico)
- Salesforce, State of the AI Connected Customer (cited in the Bain study)
- WhatsApp Business / Kantar: The State of Business Messaging (2026)
- Blip: WhatsApp Business statistics in LatAm 2026
Cristian Pereyra · Co-founder · Engineering, StudioChat
Want agents like these working for your team?
Talk to us now