We use essential cookies to make our site work. With your consent, we may also use non-essential cookies to improve user experience and analyze website traffic. By clicking “Accept,” you agree to our website's cookie use as described in our Cookie Policy and Privacy Policy. You can change your cookie settings at any time by clicking “Preferences.”
    Skip to main content
    Back to Lekta Blog[ about CUSTOMER SERVICE ]

    Customer service is stuck in the 1960s

    Sep 2, 20266 minLekta Team
    Illustration for Customer service is stuck in the 1960s

    Most industries suffer from inertia. Where we are today is usually the result of decades of small optimizations layered on top of a design that made sense once, and has quietly drifted out of date ever since. Customer service is one of the clearest examples.

    The Broken Model

    Think about what happens today when something goes wrong. You leave the app you were already using, hunt down a support number, sit through an IVR menu, re-authenticate, and wait. The system has no idea who you are, what you prefer, or why you're calling — so you explain your case again, and again, to whichever automated or human agent picks up next.

    Most companies automating customer experience have simply layered AI voice bots on top of this same outdated design. It does little to fix what actually annoys people about customer service — if anything, it makes it worse. Gartner reports that 64% of customers actively dislike AI in customer service. That number isn't really about AI. It's about AI imitating a broken process instead of replacing it.

    Meeting Customers Where They Already Are

    Look at how people actually interact with companies today, and the fix is almost self-evident. Most interactions already happen inside a mobile app. You're already logged in. The company already knows who you are and has your history and context on hand. Yet the moment something goes wrong, the experience collapses straight back to a 1960s model: find a phone number, call, authenticate again, explain yourself from scratch.

    We build the alternative: an AI assistant embedded directly in the app a customer already uses, where they're authenticated by default and the system already has the full picture.

    Instant Context Without Repetitive Interrogations

    Because the interaction happens inside your existing digital environment, the system already has the full picture. If a payment fails or a connection drops, it doesn't ask you to explain the problem — it recognizes it and proactively offers the fix.

    Helpful at the Right Moment

    Because the system understands the context behind an issue, it can offer something useful in the moment — like a credit limit increase the instant a transaction is rejected, or a bigger data plan right before a customer runs out, or a roaming package the moment they land overseas. Sales driven by timing and context, not random, cohort-driven outreach.

    Picture a real example: a payment fails, and you ask your banking app's assistant why. The system already knows who you are and which transfer you mean. No re-authentication, no re-explaining. It simply tells you the transfer exceeded your daily limit, and offers to raise the limit and reruns the payment on the spot.

    Built to Be Used, Not to Win Benchmarks

    None of this was designed to produce the smartest possible AI — the goal was an AI that customers actually want to use. Keeping a customer inside an automated flow, rather than losing them to a hang-up or an escalation, is the single biggest lever for the ROI of any AI customer service investment. No matter how capable the underlying model is, an abandoned interaction means the investment failed, and a frustrated customer is worse for retention than no automation at all.

    Since our founding in 2017, we've focused less on leaderboards and benchmarks and more on a simpler question: what's actually broken about the way customer service works today, and what would a version built around the customer look like instead? The problem is the friction baked into the workflow itself: finding the number, waiting, authenticating, re-explaining, waiting again for a resolution that could have been immediate.

    By refusing to simply bolt AI onto that legacy process, we ended up somewhere different: a customer service architecture that isn't blind to intent and context, one that we are genuinely excited to use ourselves.

    [ KEEP READING ]

    Explore more articles

    Back to blog