Be Cautious!
AI is great at speed, scale, and handling routine questions. It is terrible at the one thing that keeps customers loyal across multiple locations: human emotional intelligence and situational judgment.
The data shows that when organizations try to have AI replace humans instead of supporting them, customer experience suffers—- badly.
What Customers Are Actually Saying
- In one large consumer study, almost 1 in 5 people who used AI for customer service said it delivered no benefit at all—a failure rate four times higher than AI in other use cases.
- A global survey found 93.4% of consumers prefer talking to a human instead of AI.
- Nearly 50% said they’d cancel a service if the company forced them through AI-only customer support, and over 40% would pay more just to reach a human.
And customers are not confused about why AI is being rolled out. More than 80% believe companies use AI mainly to cut costs, not to serve them better.
Where AI Fails Customers
1. It can’t do real empathy
AI can sound polite. It still can’t feel what the customer is feeling—or respond with genuine empathy when stakes are high.
- Research tied to Zurich Insurance and Stanford shows 73% of consumers avoid businesses that don’t show empathy, and 43% actually leave them.
- 71% say AI cannot recreate genuine human connection, and 92% say they value direct human interaction more than 24/7 access.
In other words: being “always on” doesn’t beat being deeply human.
2. It struggles with complex, messy situations
Multi‑location organizations live in the world of exceptions: Local rules, weird edge cases, legacy contracts, one‑off promises, and “we did this differently at the other location.”
That’s exactly where AI breaks.
Studies and industry data show that:
- Around 75% of customers feel chatbots struggle with complex issues.
- AI systems routinely fail when the issue is ambiguous, crosses multiple policies, or sits outside the data on which they were trained.
- Even when AI supports human agents, researchers at Harvard Business School found that AI struggled with repeat complaints that required judgment, not just information.
The more nuance and context an issue requires, the more you need a human who can weigh risk, bend a rule, or make a judgment call.
3. It doesn’t build trust or relationships
Multi‑location service brands don’t just sell a product; they sell reliability and relationship.
- Research in customer behavior shows that a large share of buying decisions are driven primarily by emotions, not logic.
- Academic and field studies show customers report higher trust and loyalty after dealing with a human agent compared with AI, especially in complex or sensitive situations.
Trust is built over time, through people who remember you, understand your context, and can say, “Given everything you’ve told me, here’s what I’d do if I were you.” AI can’t do that—yet that’s exactly what keeps customers coming back to the same brand in different locations.
The Dark Side: When Companies Intentionally Make Support Hard to Reach
Here’s something even more troubling: research shows some companies deliberately design customer service to be difficult to access.
Legal scholars call this practice “sludge”—administrative barriers like endless phone trees, hidden contact information, and repetitive AI loops that are specifically designed to frustrate customers into giving up.
Why Companies Do This
The math is simple: Phone support with real humans costs $5-15 per interaction, while AI can handle inquiries for $0.50-2. If companies make support painful enough, a percentage of customers will simply give up, allowing the company to avoid the cost of resolution entirely.
The Federal Trade Commission has documented these tactics in their report “Bringing Dark Patterns to Light,” calling them “obstruction patterns”—design practices that make desired actions (like reaching a human or canceling a subscription) significantly more difficult than alternatives.
The Tactics
Research reveals specific strategies companies use:
- Hiding contact information deep in websites or eliminating phone numbers entirely
- AI programmed to close conversations rather than escalate to humans
- Endless automated loops that cycle through irrelevant options without offering human access
- Setting internal targets for reducing refunds and payouts that incentivize frontline teams to deny help
Some companies even program phone systems to disconnect after a certain hold period or require customers to use specific language to access services.
What Customers Know
Consumers aren’t fooled. Research shows:
- 76% of U.S. adults believe companies intentionally make it difficult to cancel subscriptions or reach support
- 92% say they’d switch to competitors who don’t use these manipulative practices
- The National Customer Rage Survey found growing numbers of Americans report feeling anger and even a desire for revenge after experiencing deliberately obstructed service
The Hidden Cost
While companies see short-term savings, the long-term consequences are catastrophic:
- Customer lifetime value lost: $500-5,000+ per frustrated customer
- Negative word-of-mouth reach: 10-15 people hear about each bad experience
- Social media amplification: One viral complaint reaches thousands
- Competitive vulnerability: Creates massive differentiation opportunities for competitors
For multi‑location organizations, this creates a strategic opening: While larger competitors deploy obstruction tactics to cut costs, you can differentiate simply by making human access genuinely easy.
What the Research Says Actually Works
High‑quality studies point to a consistent answer:
AI should support humans, not replace them.
- A Harvard Business School study found that when agents used AI as a copilot, they responded 22% faster and customer sentiment improved—especially for less experienced agents.
- But when responses were too fast and scripted, customers assumed they were still talking to a bot… and satisfaction dropped.
- A major management report found 95% of generative AI pilots at companies are failing, not because the tech is useless, but because organizations are trying to make AI do work that still requires human judgment.
- A Harvard Business Review–linked survey shows only 6% of companies trust AI to run core processes on its own. Most are only comfortable using it for routine, low‑risk tasks.
So the winning play for multi‑location organizations is clear:
- Let AI handle the routine: FAQs, simple updates, order status, basic reschedules.
- Let humans own the relationship: complex issues, gray areas, negotiations, escalations, and anything involving emotion or risk.
- Make human access easy and transparent—it’s now a competitive differentiator.
For customer‑facing multi‑location organizations, the research is brutally simple:
AI can automate tasks, but it cannot replace the human emotional intelligence and situational judgment required to build trust and loyalty across locations.
When leaders forget that—or, worse, deliberately obstruct access to humans to save money—customers don’t just get frustrated with the bot. They lose confidence in the brand—no matter which location they visit.
And in a world where 92% of consumers will switch to competitors who respect them enough to offer real human support, that’s not a cost‑saving strategy.
It’s a business‑ending one.
