There’s an idea gaining traction in customer experience (CX): that the fastest path to customers’ acceptance of artificial intelligence (AI) is simply telling them it’s there.
Not burying the disclosure or softening it with a vague persona name and hoping no one asks. Actually telling customers upfront that they’re talking to an AI and giving them an immediate way out if they’d prefer a human.
The result, according to practitioners who attended our latest Virtual Executive Roundtable who’ve deployed this at scale? The vast majority of customers stay and engage.
In an era of debate about whether AI can match human service quality, organizations that lead with full transparency find that only a small fraction of customers opt out. And those who stay report customer satisfaction (CSAT) scores that, while not yet at parity with human agents, are closing the gap fast. Some deployments are seeing AI-handled interactions score 93% compared to 98% for humans.
What Actually Frustrates Customers
Most organizations approach AI deployment with a reasonable-sounding concern: if we tell customers they’re talking to a bot, they’ll disengage, they’ll demand a human, or they’ll leave frustrated. This instinct is understandable.
A Gartner survey found that 61% of consumers want the ability to toggle AI interactions on or off, reinforcing that customers don’t reject AI; they just want agency over it.
What actually frustrates customers isn’t AI; it’s deception. It’s filling out a form, expecting a live agent, only to discover mid-conversation that you’ve been talking to a bot the whole time. It’s being transferred without context, left waiting, or handed off to a human who has no idea what was already discussed.
Transparency short-circuits all of that. When customers know from the start what they’re dealing with, and know they can leave if they want to, they calibrate their expectations accordingly. And when the AI delivers, they’re satisfied.
What Transparency Looks Like
In practice, the organizations seeing these results aren’t just slapping a disclaimer at the bottom of a chat window.
They’re doing a few things deliberately:
- A named AI persona. Not “our virtual assistant,” an actual character with a name that sets clear expectations. This gives customers something concrete to interact with, rather than an ambiguous blob of automation.
- An immediate opt-out. From the very first message, without a maze of options or having to express frustration to trigger a transfer. Just a clear path to a human, available at any point.
- Continuity through handoff. When a customer does choose to transfer, the human agent receives the full conversation history. Nothing is repeated, nothing is lost. The handoff feels seamless rather than punitive.
Salesforce’s 7th edition State of the Connected Customer report, surveying 16,500+ individuals, found that trust in businesses to use AI ethically dropped from 58% in 2023 to just 42% in 2024, a 16-point decline in a single year.
From the same Salesforce report: 89% of customers say it’s important to know when they’re communicating with AI rather than a human, yet most brands still don’t consistently disclose it.
The low opt-out rates practitioners are seeing aren’t a sign that customers are trapped. They’re a sign that the experience is good enough that customers don’t feel the need to leave.
Some Interactions Shouldn’t Be Handled by AI
These findings challenge something deeper than an operational assumption. They challenge a C-suite narrative that has framed AI deployment as an inherently delicate act of customer management, to be introduced gradually, carefully disguised, and constantly hedged.
That framing has it backward. Customers are not fragile or universally hostile to AI. What they are is perceptive. They know when they’re being misled, feel the friction when a bot pretends to be something it isn’t, and they reward honesty with engagement.
There’s also an important exception worth noting: some interactions should never be handled by AI, full stop. Pet loss. Genuine financial hardship. Emotional escalations. These are human moments, and the organizations getting this right have been explicit about drawing that line. Transparency isn’t just about telling customers what they’re talking to; it’s about knowing, as an organization, where your AI should never go.
Are You Solving for the Wrong Problem?
If your organization is still in the pilot phase, debating how visible to make your AI deployment and nervous about customer backlash, the data suggests you may be solving the wrong problem. The question isn’t how to introduce AI without customers noticing. The question is how to build an AI experience good enough that customers don’t mind.
Transparency is the foundation of that. Not because it’s the right thing to do (though it is), but because it’s what works. It sets accurate expectations, builds trust, and creates the conditions under which customers will give your AI a fair chance — and, as practitioners are finding, most of them will decide it earns one.
The brands that will lead in AI-powered CX over the next few years won’t be the ones that hid their AI the longest. They’ll be the ones who trusted their customers enough to be honest and built something worth being honest about.
These are exactly the kinds of candid, leader-level conversations happening inside our Virtual Executive Roundtables. If you want to hear how your peers are thinking through the opportunities and tradeoffs ahead, and contribute your own perspective to the discussion, join us at a future Virtual Executive Roundtable. It is a chance to step into a smaller, high-value conversation with fellow CX leaders navigating many of the same questions in real time.
