Step 1: Meeting your customers where they are
Connecting a dozen regional and functional mailboxes, 300+ Slack Connect channels across Enterprise Grid, Microsoft Teams, and programmatic alerting pipelines can be an exercise in IT coordination and metadata hygiene. This guide helps you navigate what’s possible so you can start off on the right foot.
What good looks like
You’ll want to see every entry point landing in Plain with its routing metadata intact, and all IT, DNS, and Slack admin approvals cleared ahead of your cutover window:
Multi-mailbox email: Your primary address and all alternate addresses have their DKIM and return-path DNS records verified, forwarding rules staged in Google Workspace or Microsoft 365, and replies automatically go out from the exact address the customer emailed
Bulk Slack Connect governance: Your Slack workspace (or Enterprise Grid workspaces) is connected, agents have completed Sign in to reply, and auto-join rules + Bulk join rules are pulling your
#ext-*/#cust-*channels in automatically as Customer-typed channels linked to their respective AccountsProgrammatic & MS Teams intake: Any in-app support forms, incident pipelines, or webhook integrations calling
createThreadpass acustomerIdentifier(and anaccountIdentifieronce the account exists) on creation
Channel architecture & governance decisions
For a larger support org, channel setup is mostly about avoiding three operational bottlenecks:
1. Clear IT, DNS, and Slack OAuth handoffs on day one
Unlike a startup where one person holds admin keys to everything, enterprise channel setup spans Google Workspace / Exchange admins (forwarding rules and distribution groups), DNS/Cloudflare admins (DKIM TXT and return-path CNAME records), and Slack workspace owners (OAuth installation). Send these requests out early and build 2-3 days of cushion into your cutover plan so you aren't blocked waiting on admin configuration.
2. Use alternate email addresses as deterministic routing keys
Instead of funneling every department into a single support@ inbox and relying on keyword guessing to split them back apart, register your functional and product addresses (billing@, security@, compute@, emea-support@) as alternate addresses.
That enables two things:
Deterministic routing: Your workflows can branch directly on
Support email equals compute@yourcompany.comAutomatic reply-from matching: When an agent replies to a thread that was sent to
billing@(even ifsupport@was CC'd or BCC'd), Plain automatically sends the outbound reply frombilling@
3. Automate Slack Connect provisioning (auto-join + account linking)
When your CSMs or onboarding automations spin up several new external Slack Connect channels a week, manual channel configuration can become quite difficult to manage. You want two rules working together:
Auto-join rules + naming conventions: Create an auto-join rule for each naming convention you use (a prefix, a suffix, or both, such as
ext-,cust-,customer-,prospect-) so the Plain bot joins new public customer channels automatically. By default, Plain types channels that start withext-,cus-,cust-,customer-, orprospect-as Customer, as well as every Slack Connect channel in your Slack workspace. Channels that match neither a rule nor our default logic (like internal#eng-*or#inc-*channels) default to Discussion channels, where messages never create customer threads.
Channel <> Account linking: Plain allows you to link Slack channels to Accounts — this sets an explicit rule that every thread from the
ext-acmechannel, for example, inherits theAcmeaccount’s tier and SLA. This is important as it will force the proper thread<>account association even when an external contractor or agency partner, who often don’t have emails that match the account domain, creates a thread in the channel.
How to implement at scale
New to Plain's email and Slack setup? Step 1: Connecting your primary support channels in our small-team series walks through it click by click.
1. Multi-mailbox email & multi-domain caveats
Email setup in Settings → Channels → Email separates receiving (forwarding each mailbox to its Postmark inbound address, something@inbound.postmarkapp.com) from sending (verifying one TXT DKIM record and one CNAME return-path record per domain). Setting up email also lets you set up the Customer Inbox.
Rather than repeating basic DNS click-paths (covered in our receiving and sending docs), here are the mechanics that matter when you're configuring 5–25 mailboxes across a larger org:
Set your primary address first, then add alternates: Your primary address must match your workspace domain and acts as the fallback
Fromaddress for customer notifications. Once saved, add your functional/regional addresses (billing@,security@,emea@) under Settings → Emails (click Add email and fill in the Add alternate email address modal) and wire each mailbox's admin-level forwarding rule in Google Workspace or Microsoft 365."This domain is already taken by a different workspace": Large orgs commonly hit this error because a team previously spun up a trial, sandbox, or departmental workspace on the same corporate domain. Deleting your support email and trying to re-add another one on the same domain will also typically throw this error. If you run into this, please ping Plain support in Slack or at help@plain.com. We can either delete an abandoned trial workspace or enable domain sharing across multiple Plain workspaces.
Alternate addresses on a secondary domain (acquisitions / multi-brand): You can add alternate support address on a different domain from your workspace as well. Add
support@acquired-brand.comunder Alternate emails and Plain shows it as Pending verification, along with the DKIM TXT and return-path CNAME records to add for that domain. Once they verify, the address can send. Addresses on your primary domain need no extra DNS records.Microsoft 365 distribution groups vs. shared mailboxes: Exchange transport/mail-flow rules don't reliably forward distribution groups to external inbound addresses. If
billing@orsecurity@is currently an M365 distribution group, convert it to a shared mailbox (or configure group-level forwarding directly) before pointing it at Plain.Multi-address deduplication & suppression hygiene: If a customer emails
billing@and CCssupport@(or forwards through an Exchange distribution header), Plain resolves the headers and creates a single thread, not two. Out-of-office replies are automatically filtered so they don't reopen closed threads, and hard bounces/spam complaints land in Email suppressions under Settings → Channels → Email.
2. Rolling out Slack Connect across hundreds of channels
Plain's Slack integration operates at three levels: Workspace (OAuth installs the Plain bot), Individual/Personal (each agent connects their Slack identity under Settings → Slack → Sign in to reply so replies post under their own name and avatar instead of the bot), and Channel (the bot must be in the channel to see or post messages).
To onboard a large backlog of customer channels seamlessly, follow these 3 steps:
Connect workspace(s) and enforce agent sign-in: Connect your Slack workspace under Settings → Slack (on Slack Enterprise Grid, connect one integration per Grid workspace that hosts customer channels). Have every frontline and T2 agent complete Sign in to reply during onboarding so outbound replies don't appear from a generic bot.
Configure auto-join rules, then run Bulk join rules: Go to Settings → Slack, pick your workspace, and open Customer channels. In the Bulk join rules card, click Add rule, choose Channels with prefix, and add each prefix (
ext-,cust-,customer-), then use Bulk join rules to pull the Plain bot into your existing public customer channels in one shot (archived channels are skipped automatically). A rule stays staged until you click Save, and Save joins the channels. For private Slack Connect channels, Slack's security model requires inviting the bot explicitly (/invite @Plainor via your Slack channel-provisioning script).Pick the right ingestion mode per channel tier:
AI-based (recommended — top-level Slack threads are grouped semantically, based on the similarity of content of the messages) or Time-based (the default — top-level Slack threads are grouped into a single Plain thread based on time + author) for standard customer channels
One-to-one when every top-level Slack thread should map 1:1 to a Plain thread
Manual emoji (e.g. 👀, scoped to Plain users or Anyone) for high-chatter channels where you only want specific messages escalated into the ticket queue
API-only when your own bot or triage layer sits in front of Slack and creates Plain threads programmatically
3. Programmatic intake (createThread on the API channel) and MS Teams
When threads are created by an in-app support form, a headless portal, or an automated monitoring/incident pipeline, they enter Plain on the API support channel via createThread.
At scale, never create an API thread with only an email address if your backend knows which workspace or product the user is in. Pass both customerIdentifier and accountIdentifier (plus an idempotent externalId for correlation) in the createThread call so the thread skips domain guessing and lands in your workflows with the proper Account and Tier attached:
mutation { createThread(input: { customerIdentifier: { externalId: "user_8821" } accountIdentifier: { externalId: "acme-prod" } externalId: "incident-2026-10-13-0442" title: "Compute us-east-1: p99 latency above threshold" }) { thread { id } } }
Verification checklist
Alternate address round-trip: Send an email to an alternate address like
billing@yourcompany.com. Confirm the thread lands in Plain and that replying from Plain delivers frombilling@yourcompany.comwith the sending domain's DKIM and return-path records showing as verified in Settings → Channels → Email.Slack Connect + personal identity check: Post from an external Slack account in a Customer-typed channel (or use Create a test channel on your Slack workspace page under Settings → Channels → Slack), reply from Plain, and confirm the reply posts in Slack under your personal Slack profile.
API thread opens: Fire a test
createThreadcall with a customer email and confirm a thread opens.
Troubleshooting/FAQ
Inbound email arrives on an alternate address, but outbound replies fail or land in spam.
Every sending domain requires verified DKIM (TXT) and return-path (CNAME) records, and Plain automatically suppresses addresses that previously hard-bounced or reported spam. Here are the two causes:
DNS records are unverified (or proxied in Cloudflare)
If either DNS record is unverified — or if your Cloudflare CNAME is set to Orange Cloud (Proxied) instead of DNS only — recipient servers will reject or spam-folder outbound replies.
Try this: Open Settings → Channels → Email and verify both records show green for that domain. If they do, walk through the deliverability troubleshooting guide.
The recipient is on your workspace suppression list
Addresses that hard-bounce or mark an email as spam are added to your suppression list to protect your domain reputation.
Try this: Open Settings → Channels → Email, find Email suppressions, and check the customer's address. If it's a hard bounce, click Reactivate email address once their mailbox issue is resolved. A suppression that followed a spam complaint can't be removed.
Emails sent to a Microsoft 365 distribution group aren't forwarding into Plain.
Exchange mail-flow (transport) rules don't reliably match distribution groups the way they match shared or user mailboxes.
Try this: Convert the distribution group to a shared mailbox in Microsoft 365 (or configure forwarding directly on the group object) and point its forwarding rule at your Plain Postmark inbound address.
A message in a customer Slack Connect channel didn't create a thread in Plain.
A Slack message only opens a thread when the Plain bot is in the channel, the channel is typed Customer, the sender is external, and the ingestion mode allows it. Check these five culprits in order:
The Plain bot hasn't been invited to a private Slack Connect channel
Auto-join rules can only join public channels automatically; private Slack Connect channels require an explicit invite.
Try this: Run
/invite @Plaininside the private channel (or add the bot via your channel-provisioning automation) and confirm it appears under Settings → Channels → Slack.
The channel defaulted to Discussion or Disabled
Channels that aren't Slack Connect channels and don't match your auto-join rules or the standard customer prefixes (
ext-,cus-,cust-,customer-,prospect-) default to Discussion channels, which never create customer threads.Try this: Open the channel in Settings → Channels → Slack and flip its type to Customer.
The test message was sent by someone on your workspace domain (or a Plain user)
Plain treats messages from your own email domain or from registered Plain users as internal team chatter, so they never open customer threads.
Try this: Test from an external Slack account outside your corporate domain, or click Create a test channel in the Slack setup wizard.
Ingestion mode is set to Manual emoji or API-only
Manual emoji waits for the configured reaction, and API-only disables automatic ingestion so your own service can call
createThread.Try this: Check the channel's ingestion mode in Plain. If it's Manual emoji, react with the configured emoji (and set the audience to Anyone if customers should be able to trigger it).
Time-based or AI grouping appended the message to an already-open thread
Time-based grouping adds a message to the existing thread if the same person wrote within 2 hours, or anyone else within 1 hour. AI-based grouping decides whether a message continues the previous thread or starts a new one based on the similarity of the content of messages.
Try this: Check the most recent open thread from that Slack channel in Plain to see if the message was grouped there.