A Small Studio Saved 15 Hours a Week by Answering Email in 14 Seconds

A small studio does not lose its evenings to strategy. It loses them to the inbox — the freeze request before class, the billing confusion after it, the message that sat for a day because everyone was teaching. Kohr Method's story is about that inbox. It is also a vendor story, and it is labelled that way throughout.

Source and label: the figures below come from a case study published by Momence, the platform that sells the AI Inbox described. They are Momence's account of a named customer, Kohr Method, a pilates and wellness studio run by co-owners Emily and Prima. The numbers are claimed by the vendor, not independently audited.

The bottleneck: good classes, slow replies

On Momence's account, customer inquiries had been waiting 24 to 48 hours, and some actions — membership freezes and cancellations handled through a form reviewed weekly — could take a week or more. Refund requests and billing confusion then needed untangling. For a studio, that delay is not an abstract service metric. The minutes before class are when a message matters most, and a slow answer turns a routine request into frustration that walks into the room.

Note the shape of the work. Most messages were routine: a freeze, a cancellation, a policy question, a class enquiry. The answers already existed in the studio's own policies and class descriptions. What was scarce was not knowledge. It was attention at the right moment, from a tiny team whose real work happens in person, in a room, on a schedule.

What Momence says the AI Inbox did

Momence says its AI Inbox was given the studio's brand voice and instructions, drawing on customer data, policies and class descriptions already held in the platform. Routine requests, including freezes and cancellations that once waited for a weekly review, came to be handled in seconds. The vendor quotes response times falling from 24–48 hours to 10–14 seconds, a saving of more than 15 hours a week, and about $2,000 a month saved by consolidating support, including phasing out a virtual assistant arrangement.

Hold those numbers at the right distance. Fifteen hours a week is a large claim for a small studio, and the vendor does not publish the timesheets, message volumes or wage assumptions behind it. The $2,000 a month appears to be a consolidation saving — support brought back in-house — rather than cash found in the inbox itself. None of that makes the story useless. It makes it a vendor account to test, not a benchmark to copy into your plan.

The part worth copying is the division of labour

The quoted line from the studio in the vendor piece is about focus: the team spending its energy connecting with clients and improving the in-person experience. Whether or not the hours are exact, that division is the right test. Routine replies were absorbed by the machine. Scarce attention went back into the room, where a studio actually earns its money and keeps its members. Attention is the budget; the inbox was spending it at the wrong hour.

Compare that with the failure pattern in the MIT article, where tools are bought for transformation and the routine layer is left untouched. A studio does not need transformation first. It needs the freeze request answered before class and the billing question closed without a week of friction. Small, named, repeatable work is where a first automation earns trust.

What to check before you put an assistant in front of your customers

Start with your own wait times. Sample fifty messages: how long did each wait, how many were routine, how many needed a person? Then read your policies the way a machine would. If the freeze rule, the cancellation rule and the class descriptions are not written clearly in one place, an inbox assistant will improvise, confidently and at speed — which is the Air Canada failure in miniature, covered later this week. Write the answers first. Decide which requests the assistant may complete, which it drafts for approval, and which it must hand over immediately. Keep the handover visible to staff, not buried in a log nobody reads.

Finally, train the people who supervise it. Someone has to own the voice, the exceptions and the weekly check of what was said in your name. That staff training point is the bridge to our consulting services: the tool is the easy purchase; the operating habit is the work.

The reply-speed number is not the prize

It is tempting to chase the 10-to-14-second figure, because it is the easiest number to beat a competitor with. Resist it. A reply in fourteen seconds that misunderstands a freeze, invents a policy or misses the one message that needed a person before class is worse than a correct reply in an hour. Measure three things instead, on your own fifty-message sample: how many answers were right without a human touching them, how many a person had to repair, and how many should never have been answered by a machine at all. The third number protects trust, the second prices supervision, and only the first is a saving. A studio that reports those three numbers honestly knows more about its inbox than a vendor slide claiming fifteen hours — including, possibly, the studio in the vendor slide.

Start with the fifty-message sample this week, before any purchase. It costs one quiet hour, and it converts a vendor's fifteen-hour promise into your own three numbers — right, repaired, and never-the-machine's — which is the only version you can safely act on.

Tell Ada how long a customer message waits on a busy day. She will map the fix — what the assistant should answer, what a person should keep — in your free report. Start at /audit/. Bring your real wait times; leave the vendor's 15 hours as their claim, not your target.

Source: Momence blog, How Kohr Method reclaimed 15+ hours a week with AI Inbox, 4 July 2025 (vendor-sourced; all figures attributed to Momence and not independently audited).