[ MSP, APIs & integrations ]

Customer notifications through Microsoft Teams, Slack or email

Choose a notification channel and prove the difference between attempted send, platform acceptance, receipt and human acknowledgment.

By the EmailAuth research team · 3 min read

[ Key points ]

  • Choose a notification channel and prove the difference between attempted send, platform acceptance, receipt and human acknowledgment.
  • Platform acceptance, receipt and human responsibility are different events.
On this page

An authentication alert is useful only when the right person can act on it before it becomes stale. Design notification delivery around an accountable destination and an expiry rule, not a single sent flag. Teams, Slack and email have different interfaces, permissions and evidence of delivery.

Choose a supported channel path

Microsoft's current incoming-webhook guide states that connector-based Office 365 Incoming Webhooks were retired in May 2026 and directs new webhook posting through the Workflows app. Verify the current supported route rather than deploying an old connector tutorial. The guide also notes that workflows belong to users, not the team or channel, so assign co-owners and test continuity when an owner leaves.

For Slack, chat.postMessage is a documented posting method with permission and rate-limit requirements. Verify the returned application result and destination identifier; an HTTP response alone is not a human acknowledgment. Respect rate-limit responses and keep retries bounded.

For email, RFC 5321 distinguishes SMTP responsibility and subsequent delivery handling. Acceptance by a server does not prove inbox placement or that a person read the message. Keep bounced or delayed delivery separate from a successful handoff, and never claim that an authentication pass guarantees receipt.

Minimize and route the message

Include customer, affected service, observation time, consequence, confidence, incident reference and the next action. Keep full reports and sensitive operational details behind a suitably restricted evidence link. Confirm channel membership and email recipients against the approved customer map. Do not reuse a convenient shared channel across tenants without an explicit access decision.

Use an episode identifier so repeats update or reference the same incident. Separate notification severity from raw report volume. A burst of repeated observations should not drown out a new critical service failure. Acknowledge ownership in the service workflow rather than assuming a chat reaction always means technical resolution.

Measure four distinct times

Record attempted-send time, platform-acceptance time, independently observed receipt time where available, and expiry. Human acknowledgment is a fifth, separate event. If receipt cannot be observed, record it as unknown rather than inventing a timestamp.

The written hypothetical timing example uses a common send time of zero seconds and acceptance at one second. Fictional Teams receipt occurs at nine seconds, Slack at four and email at forty. The illustrative expiry is thirty seconds. Therefore email is marked expired and escalated even though it eventually arrives. All human acknowledgments remain false. These are invented clock values to demonstrate logic, not platform benchmarks or real delivery observations.

Read those invented times as a written timeline, not an event-loop run. At second 30, Teams and Slack have stipulated receipt while email has not, so the proposed rule marks email late and requests an approved escalation. At second 40, its stipulated late receipt does not erase the earlier missed deadline. Receipt is an assumed observation field, not a claim that every channel exposes a delivery-receipt API.

No simulated transport,clock dispatch or external message was executed for this revision. A real test needs an authorized destination,actual platform response and a defined observation method for each channel. If receipt cannot be independently established, keep it unknown; do not use posting success as its timestamp.

Plan escalation without multiplying messages

Agree who owns an expired notification and which fallback is permitted. A fallback message is a new communication with its own recipient and content scope. Do not automatically post private evidence somewhere broader when the first channel fails. Keep delivery retries separate from escalation policy.

Test revoked permissions, a removed destination, throttling, an orphaned Teams workflow and a delayed email. Review notification routes when customer contacts or channel membership changes. These checks support an accountable process; they do not establish EmailAuth channel integrations or guaranteed response times.

[ Takeaway ]

Platform acceptance, receipt and human responsibility are different events.

[ By role ]

What this means for you

Choose your role for practical next steps.

IT / Mail Admin

Instrument each delivery stage

Channel interfaces expose different evidence.

Next steps

  1. Retain acceptance separately from receipt.
  2. Test throttling and missing destinations.
  3. Use current Teams workflows.

CISO / SecOps

Protect recipient scope

Delivery recovery must not broaden data exposure.

Next steps

  1. Verify channel membership.
  2. Approve fallback recipients.
  3. Audit evidence-link permissions.

Founder / Non-technical

Define response accountability

A sent message is not an accepted operational responsibility.

Next steps

  1. Agree expiry and escalation.
  2. Assign channel owners.
  3. Review unacknowledged critical incidents.

[ FAQs ]

Frequently asked questions

Should new Teams alerts use an old connector tutorial?

No. The cited current Microsoft guide directs webhook posting through Workflows after connector retirement.

Why assign a Teams workflow co-owner?

Workflows are user-owned; co-ownership helps avoid an orphaned automation when an owner leaves.

Does Slack posting success prove someone read it?

No. Platform acceptance and human acknowledgment are separate observations.

Does SMTP acceptance prove inbox delivery?

No. It does not establish inbox placement or human receipt.

Are the sample channel times measured?

No. They are invented values in a written timeline, not platform measurements or an executed model.

What if receipt cannot be observed?

Record it as unknown and apply the agreed expiry and escalation policy.

Can a failed alert be sent to any fallback channel?

No. Fallback recipients and content must remain within the approved communication scope.

What belongs in the alert body?

A minimal actionable summary, incident reference and restricted evidence link, not unnecessary full message content.

[ Sources ]

Sources

  1. Microsoft: Create incoming webhooks using Workflows

    Sections Important retirement notice · Limitations · Create Webhooks using Workflows

  2. Slack: chat.postMessage

    Sections Usage info · Response · Rate limiting · Channel membership

  3. RFC 5321: Simple Mail Transfer Protocol

    Sections §3.3 · §4.2.1 · §4.2.5 · §5.1 · §6.1

[ Glossary ]

Glossary

Notification expiry
The point after which an alert needs a different action rather than ordinary delivery retry.
Human acknowledgment
An explicit indication that a responsible person has accepted the notification or assigned task.

Enterprise provenance

Why top financial institutions and governments retired their email security dashboards.

Traditional email authentication platforms sell you a dashboard and ask your engineers to do the labour. In high-stakes environments, manual oversight means delayed enforcement, configuration drift and unmonitored shadow domains.

EmailAuth is the agentic email provenance platform. Our agents enforce p=reject, keep third-party senders aligned and move your domains to full RFC 9989 compliance, the new DMARC standard,, with human-in-the-loop auditability.

Governments Large enterprises Tier-1 financial institutions

SaaS · Dedicated tenancy · Air-gapped on-prem · BYOK encryption

Knowledge Base