Here's a support interaction your customers know by heart. They email "where is my order?" Twenty hours later, a reply asks for the order number — which was in the subject line. They send it again. Another day passes. Total elapsed time to receive one tracking link: three days. The tracking link was a 30-second answer the whole time.
Email is a fine medium for receipts and contracts. It is a terrible medium for conversations, and post-sale support is conversations. WISMO — "where is my order" — runs 40 to 60 percent of ticket volume for a typical DTC brand, and every one of those tickets is a half-minute answer trapped in a 24-hour medium.
The way out is moving the conversational half of support to a channel built for conversation, without breaking anything in transit. Here's the migration plan we've seen work, in four phases over about six weeks.
Why a switch fails and a migration works
The tempting version is the cutover: announce a Telegram support channel, put "we no longer offer email support" in the footer, done. What actually happens: customers mid-thread get orphaned, the help page sends people to a channel they've never used, and your first week is spent apologizing.
The version that works runs both channels in parallel and demotes email gradually, on evidence. Nothing gets turned off until the numbers say it's safe.
Phase 1 (week 1): Open the instant doors
Put chat entry points where customers already are. No announcement yet — just doors:
- The order-status page and your site, via the widget. This is genuinely a two-line job:
<script async src="https://cdn.lane.chat/widget.js"
data-lane-id="wgt_xxxxxxxxxxxx"></script>- The order confirmation email. One line above the footer: "Questions about your order? Message us on Telegram — t.me/YourBrandBot. We're much faster there."
- The packing slip. A QR code straight to the bot. Someone holding your product with a question is the highest-intent support contact you will ever get; don't route them to a webform.
Why Telegram as the anchor channel: your customer already has it, there is nothing to install, the conversation history survives phone upgrades, and — unlike email — a reply actually arrives as a push notification instead of landing in a promotions tab.
Email stays exactly as it is. You've added doors, removed nothing.
Phase 2 (weeks 2-3): Route by intent, not by decree
Not everything belongs in chat. Split the traffic honestly:
Chat gets the conversations: WISMO, return initiation, sizing and product questions, "can I change my address" — anything with back-and-forth in it.
Email keeps the records: receipts and invoices, warranty claims that need documents-of-record, legal, chargebacks. Anything a customer might need to forward to a bank someday.
The routing mechanism is soft. When a WISMO email comes in, answer it fully in email — then add one line: "Next time, message us on Telegram and you'll have this in seconds." No forced redirects, no "we don't handle that here." You're advertising the fast lane, not tearing up the road.
By the end of week three, watch what your customers do. New WISMO contacts should already be tilting toward chat, unprompted. People aren't loyal to email; they're loyal to whatever answered fast last time.
Phase 3 (weeks 3-4): Put AI on the first line
This is where the economics flip. An instant channel with humans behind it is just a faster treadmill — you've promised speed and now have to staff it. The instant channel earns its keep when the first line answers itself:
- Load the knowledge base: shipping matrix by region, returns policy, product FAQ. Every AI answer should cite its source doc, so your team can audit answers instead of hoping.
- Wire order-status lookups so "where is my order" gets a real tracking answer, not a policy paragraph.
- Set the human-takeover triggers: refund demands, visible frustration, or two low-confidence answers in a row — a person steps into the same thread.
The pattern that emerges: roughly the same handful of questions all day, answered in seconds, with your team touching only the conversations that actually need judgment. That's the treadmill turned off.
Phase 4 (weeks 5-6): Measure, then demote
Now compare the channels on four numbers: median first response, time to resolution, repeat-contact rate (did they have to ask twice?), and share of conversations resolved without a human.
If chat is winning — it will be — demote email visibly: chat goes first on the help page and in the footer, email drops to "for invoices and formal requests." Do not kill email. It remains the paper-trail channel, the fallback, and the place chargebacks live. Demoted, not deleted.
The four ways this goes wrong
- Forcing an app install. Brand-app-only chat fails because nobody installs an app to ask about socks. Telegram and WhatsApp win precisely because they're already on the phone. Meet customers where they are; that's the entire trick.
- The silo tool. If chat lives in a separate console nobody opens, response times quietly rot and you've rebuilt email with extra steps. Mirror conversations into the messenger group your team already watches, or at minimum into one shared inbox with everything else.
- The timezone gap. An instant channel raises expectations around the clock, and your customers' evening is your team's 3 a.m. AI coverage plus an honest off-hours note — "a human joins after 9am ET" — beats silence every time.
- Migrating outbound too early. Shipping notifications and proactive pushes can move to Telegram too — later. That's phase five. Bundle it into the launch and you'll debug two systems at once.
What good looks like
Six weeks in, the brands that run this playbook see the same shape: roughly half of WISMO never touches a human, median first response drops from most-of-a-day to seconds, email volume falls by half or more — and nothing customer-facing broke on the way, because nothing was ripped out. The conversations moved to a conversational channel. The records stayed where records belong.