If you’ve ever tried to get a real answer from a telecom operator or a travel eSIM app when something went wrong abroad, you already know the pattern: a chatbot greets you instantly, cycles through the same three suggested articles regardless of what you typed, and if your problem doesn’t match one of those articles, you’re often looking at a 24-to-48-hour wait for an actual person — while you’re standing at an airport with no data. This isn’t bad luck or a few companies cutting corners. It’s the predictable result of a cost structure that most of the industry has quietly decided to accept, and it’s worth understanding exactly why, because the economics explain almost everything about the customer experience you get.
The math that drives the decision
Here’s the number that explains the entire industry: a human-handled support ticket costs somewhere between $8 and $25 depending on channel and complexity, with phone support at the expensive end and email or chat somewhat cheaper. Forrester and SQM benchmarks put the average around $13.50 per contact. An AI chatbot resolving the same routine query costs roughly $0.10 to $1.50. That’s a 12x to 24x cost differential on the exact same interaction, before you even factor in staffing headaches.
And the staffing headaches are real. Call center attrition runs 40–45% annually industry-wide, and first-year attrition for new agents is closer to 65–70%. Replacing a single agent costs $10,000 to $20,000 in direct expenses, or up to $46,000 once you count lost productivity during ramp-up. A support team handling just 5,000 tickets a month is burning roughly $67,500 monthly in human resolution costs alone, before software, real estate, or management overhead. Set against that backdrop, a chatbot that resolves 80% of routine inquiries for pennies per interaction isn’t a corner-cutting decision — from a pure spreadsheet perspective, it’s close to the obvious one.
Why this hits travel eSIM particularly hard
Travel connectivity apps are a near-perfect case study in why this economic pressure produces the exact experience travelers complain about most. Take Airalo, the largest global player in the category — not to single it out unfairly, but because its public review record is genuinely representative of the category’s structural problem, not an outlier. Airalo currently sits around 2.1 stars on PissedConsumer, with close to 70% of reviews unfavorable. Look through the complaint patterns and one theme dominates: an AI chatbot handles first contact 24/7, and when a problem needs a human — a failed installation, a billing dispute, a refund — the wait runs 24 to 48 hours. There’s no phone line at all; support runs through chat, email, and WhatsApp only. One reviewer summed up the frustration precisely: they knew any response would come from an AI, “as many people have pointed out, no human.”
The structural reason this pattern is so common in travel eSIM specifically, more than in most other consumer categories, is that the business model doesn’t require an ongoing relationship. A traveler buys a data package for a two-week trip and, in the ordinary case, never contacts the company again either way. That means the commercial incentive to invest in support infrastructure is weaker than it would be for, say, a subscription service where a bad support experience directly threatens recurring revenue. Combine that with venture-backed growth targets that reward customer acquisition volume over support quality, and chatbot-first support isn’t a bug in the model — it’s a fairly rational output of it.
Worth being clear, though: this isn’t a travel-eSIM-specific failing. The same pattern shows up, with remarkably similar numbers, among the large incumbent mobile network operators that have been in business for decades. The American Customer Satisfaction Index’s 2026 telecommunications study puts the industry-wide score at just 77 out of 100 — one of the weaker showings across all industries ACSI tracks, trailing well behind banking and software specifically because of “billing complexity, outages, and limited customer choice.” More tellingly, value-oriented MVNOs consistently and measurably outperform the large legacy carriers on satisfaction, year after year, with a 17-point gap between the best and worst performers in the 2026 study alone. The pattern isn’t “new travel apps are worse than established telecoms.” It’s “any operator whose scale and cost pressure push it toward automation-first support tends to underperform smaller, more service-focused competitors” — regardless of whether that operator has been around for two years or twenty.
What good human support actually costs — and why it’s worth it anyway
None of this means chatbots are bad technology. A well-built chatbot genuinely resolves simple, well-defined questions faster than a human could, and there’s no reason to make a customer wait for a person to answer something a knowledge base article already covers clearly. The problem isn’t chatbots existing. It’s chatbots being the *only* option when the issue is time-sensitive, emotionally stressful, or doesn’t fit a predefined script — which describes almost every support request that actually gets escalated, by definition, since the easy ones never reach that point.
Running real human support, available when someone is stranded without connectivity in an unfamiliar country, costs what the numbers above suggest it costs — call it somewhere in the range of $10–20 per resolved contact once you include training, turnover, and the simple fact that good agents who actually solve problems don’t work for the same low wages that produce 65% first-year attrition. That’s a real, ongoing cost line, not a marketing slogan. Choosing to carry it, instead of routing every difficult case through an AI escalation queue with a two-day wait, is a deliberate tradeoff — and it’s the specific tradeoff we’ve made at NoveSIM from the start. It’s also, frankly, one of the harder things to communicate to a customer before they need it, because nobody reads the “we have human support” line until the moment they’re standing at a foreign airport with no signal, wishing someone would just answer the phone.
The takeaway for the industry
The uncomfortable truth is that customer support quality in travel connectivity isn’t really a product feature — it’s a direct, visible readout of which companies are optimizing for unit economics on paper versus which ones are optimizing for what actually happens to a real traveler when something breaks. Both are legitimate business strategies. But travelers deserve to know which one they’re buying into before they need help, not after.