Nobody builds an app because they dream of answering emails about it. And yet: the moment you have users, you have support — whether you set it up deliberately or let it happen to you. The second option is how founders end up answering "is your app down?" messages at 23:40 with a sinking feeling.
Here's the setup that works for one person. It's not glamorous. It works.
One inbox. Seriously, one.
No chat widget, no WhatsApp, no Discord DMs, no Instagram comments. One email address, printed clearly on your site. Every channel you add is a place you can now fail to answer — and an unanswered chat widget looks far worse than no chat widget at all.
A chat bubble on the site signals "we answer in minutes." Do you? At 3 a.m.? Exactly. Email signals "we answer within a day," which is a promise a solo founder can actually keep.
Say your response time out loud
Put it on the contact page: "We answer within one business day." Then do that. An honest 24 hours beats a broken promise of 2 — customers are surprisingly patient when they know what to expect, and surprisingly furious when they don't.
Sound familiar? It's the same principle as an SLA, just informal. Committed expectations, kept. That's the whole trick, in support and in operations.
The triage rule that fits on a sticky note
When a message arrives, ask two things: is money moving wrong, and is data at risk?
- Yes to either → drop what you're doing. This is the fire lane: failed payments, double charges, users seeing someone else's data, anything security-shaped.
- No to both → it waits its turn in the one-business-day queue. Feature requests, "how do I", cosmetic bugs. All of it.
That's the entire triage system. The failure mode it prevents: spending your morning on a font complaint while a webhook backlog quietly eats renewals.
Write the answer once
After two weeks of support you'll notice the same five questions keep coming. Write your best answer to each — once, properly, with screenshots — and save them somewhere you can paste from. Personalize the first line, paste the rest.
This isn't laziness. Your fifth answer to the same question is better than your first; canned replies mean every customer gets your best version. And those five recurring questions? They're a to-do list. Each one is either a missing help paragraph on your site or a confusing screen in your app. Fix the cause and the question stops arriving.
Let AI draft, never send
Yes, use AI in support — for drafting replies, summarizing long angry threads, translating. It's genuinely good at that. But read everything before it goes out, because the one thing AI reliably lacks is knowledge of what you actually promised this specific customer. A support reply is a small contract. Sign it yourself.
Bugs: the honest choreography
When someone reports a real bug, three sentences keep the relationship intact: confirm you can reproduce it (or ask for details, once), say what happens next, and — this is the one people skip — close the loop when it's fixed. "This is fixed now, thanks for catching it" turns a bug reporter into a fan. It costs you one email.
What breaks trust isn't bugs. Every app has bugs. It's silence.
Refunds: decide the policy before you're angry
Write your refund rule now, while nobody is yelling: something like "full refund inside 14 days, no questions." Generous is cheaper than the hours of argument, and about a thousand times cheaper than a chargeback dispute. When the edge case arrives at 22:00 — and it will — you follow the written rule instead of your mood.
The part support can't fix
Here's the pattern we keep seeing: half of "support" load on AI-built apps isn't really support. It's operations wearing a support costume. The "app is slow" emails, the "my payment didn't go through" threads, the "I got a weird error" screenshots — those are monitoring gaps and unhandled failures surfacing through your inbox, one confused customer at a time.
You can answer those emails forever, or you can fix the machinery that generates them: proper monitoring, tested recovery, patched dependencies. That's the half we take over when an app comes to us — and inbox volume dropping is usually the first thing owners notice.