Twenty Spam Calls a Week, and an AI Picks Up for You
In India twenty unwanted calls a week are now the norm. An AI answers for you, asks the stranger who is calling, and rings you only if it is worth it. At what cost?
In India, the average smartphone owner fields around twenty unwanted calls a week: telemarketers, debt collectors, fake couriers, scams of every stripe. The phone, meant to connect us, has turned into a dispenser of interruptions, to the point that many people no longer answer when an unknown number lights up. It is on this weary ground that an Indian startup, Equal AI, raised thirty million dollars in June 2026, in a round led by Prosus Ventures and Tomales Bay Capital.
Its app does not merely block calls: it answers them. A synthetic voice picks up on your behalf, asks the stranger who they are and what they want, then hands you a summary while the line waits. More than a million people use it every month, three hundred and fifty thousand every day. And the trick is no longer one startup's alone: since iOS 26, Apple has baked it into the system, and Samsung and Google offer their own. The talking filter, stationed between you and your calls, is quietly becoming a default setting.
A voice that answers before you do
The mechanism comes down to an inversion. On an iPhone running iOS 26, a number missing from your contacts no longer hears your ringtone: it meets a prompt asking for a name and a reason for calling. Whatever the caller says appears live on your screen, transcribed in a few words. You read it, then choose to take the call or let it go. After thirty seconds of silence, the stranger is offered the chance to leave a message. Your contacts still ring through as usual: the filter applies only to numbers you do not know.
Equal AI goes a step further. Its voice does not just listen, it negotiates. Faced with a courier, the app offers ready-made replies, "leave the parcel at the door" or "give it to a neighbor," and reads them aloud to the caller. The assistant holds a full conversation, grasps intent, and decides on its own whether to let the call through, take a note, or hang up. You learn the call ever happened only afterward, as a summary.
In both cases the promise is the same: to stop being the first line of defense for your own phone. The chore of triage, that reflex of grabbing the handset to check whether it is worth it, moves to the machine.
A ring you have to earn
The gain is measured in interruptions avoided. A call that never lands is a meeting not broken, a meal not paused, a train of thought not snapped in two. For someone taking twenty junk approaches a week, the math is quick: the phone now rings only when a human has bothered to give a name and explain the visit.
There is a kind of autonomy in this. For years the phone set the rule: it rings, you run. The handset dictated the tempo, the caller chose the moment, the recipient endured. The filter reverses that. Now it is you who set the terms of access to your attention, and a stranger must clear a gate before reaching you. The cold caller who counted on your curiosity, or your fear of missing something, hits a door that is polite but shut.
For the elderly, so often the first targets of phone scams, the benefit is sharper still: a buffer slips between them and the manipulator and absorbs the first blow. The machine cannot be rushed or intimidated.
What the filter leaves outside
Yet not everything that matters arrives under a familiar name. The lab calling back with a result, the school reporting a sick child, the tradesman awaited for three days, the prospective employer: all legitimate calls from unknown numbers, precisely the ones the filter treats as suspects. Many callers, meeting a robotic voice, hang up without a trace, sure they have reached a machine or a scam. The important call is not blocked, it evaporates.
American consumer groups that tested Apple's feature put it plainly: the comfort is real, but it shifts the risk. You no longer miss the nuisances, you risk missing what counts. And the marketers' countermove came fast: guides already circulate teaching sales calls how to clear the gate by stating a credible reason the machine will wave through. The filter does not abolish the nuisance, it opens a fresh race between the gatekeeper and those who want around it.
Your switchboard, their server
Finally there is what you give up for the quiet. Handing your calls to an assistant means letting a third-party ear listen to every stranger who seeks you, transcribe their words, sometimes keep a record of who called, when, and why. On Apple's phones the processing happens on the device, which limits the leak. With an independent app, the exchange usually travels up to a server, and the log of your calls becomes data living somewhere other than your pocket.
The dependence runs two ways. Dependence on an intermediary that, should it break or fold, leaves you exposed again. And dependence on an automated judgment you cannot override in the moment: it is the one deciding, in a single sentence, whether a call deserves to reach you. The sorting you once did on instinct, with its errors but also its common sense, is handed to a system whose verdict is all you see.
The phone used to be a wire strung toward you, open to whoever dialed the number. Now it comes with a vestibule where an agent screens visitors before disturbing you. For many the trade is excellent: giving silence back to fractured days is well worth a little delegated vigilance. The real question is not whether we want this gatekeeper, but how far we let it decide, on its own, who has the right to reach us.