Skip to main content
Back to blog
Product Updates

Smart Response Handling: How AngelSend Sorts and Drafts Replies

By · Updated Oct 28, 2026 · 7 min read

Smart Response Handling: How AngelSend Sorts and Drafts Replies — AngelSend blog article featured image

When outreach works, replies start arriving, and they aren't all the same. Some people are interested, some have a question, some are out of office, and some want you to stop. Smart response handling is AngelSend's name for the features that sort those replies and help you answer them, with a person reviewing anything that matters.

Update, October 2026: This post has been rewritten to describe how the feature works today. An earlier version included usage statistics and beta results that we couldn't support, and they have been removed.

The Problem: One Inbox, Many Kinds of Replies

A campaign's replies come in on a single channel but call for very different actions. An out-of-office message should usually just postpone the next touch. A question about pricing needs an accurate answer. "Please remove me" has to stop all contact. "You want my colleague" is a referral, not a rejection. When all of these land in one undifferentiated inbox, the urgent replies wait behind the routine ones, and it's easy to miss something important.

The goal is not to automate every conversation. It's to separate the predictable work from the replies that need judgment, so people spend their time where it counts.

How It Works

1. Intent classification

When a reply to a campaign arrives, AngelSend analyzes it and records a detected intent. The main categories:

  • Interested: wants to learn more or talk
  • Question: asking for information before deciding
  • Objection: raising a concern such as price, timing, or fit
  • Not interested: a clear no
  • Out of office: an automatic away message
  • Wrong person: the contact isn't the right owner, sometimes with a pointer to who is

Classification is a model's judgment, not a certainty. Ambiguous or mixed replies happen, so treat the label as a routing aid rather than the final word.

2. Follow-ups stop when someone replies

Follow-up rules re-contact people who haven't replied, using triggers such as No Reply, Never Opened, or Opened But No Reply, on a delay you choose. Once a contact replies, the scheduled nudges stop, so nobody gets a "just checking in" email right after answering you.

3. Auto-reply rules draft a response

Auto-reply rules route replies by detected intent and draft a response. You can give per-intent instructions, for example how to answer a common pricing question or what to say when someone asks to be contacted next quarter, and those instructions take priority over the generic drafting guidance. Keep the instructions factual and limited to approved positioning. The model shouldn't invent prices, features, or commitments.

4. Drafts go to a review queue

By default, AI-drafted replies land in the Review queue for a person to approve, edit, or discard before anything is sent. You can turn approval off for categories you trust, but we recommend leaving it on until you've read enough drafts to know how they behave for your audience.

5. Opt-outs are honored everywhere

Recipients can opt out through the unsubscribe link in every campaign email's footer. An unsubscribe suppresses that address across every campaign, not just the one it came from. Under CAN-SPAM, opt-out requests must be honored within 10 business days, and Gmail and Yahoo's February 2024 bulk-sender rules expect unsubscribes to take effect within two days. Suppression should be immediate and organization-wide.

Test Before You Launch

The Automation step of the campaign builder includes a test card that dry-runs your follow-up rules and auto-reply configuration on a campaign that hasn't sent anything yet. It shows a readiness checklist (rules, auto-reply configuration, connected mailbox, booking link), the follow-ups that would go out, and how sample replies of each type would be classified and answered. Nothing is queued, drafted, or sent. It's the quickest way to catch an automation that looks switched on but has nothing behind it.

A Conservative Way to Configure It

  1. Start with review on for everything. Read the drafts for a week or two before relaxing anything.
  2. Write instructions per intent. Say what's true and approved, link to real pages, and say what the draft must never promise.
  3. Send sensitive topics to a human. Contract terms, custom pricing, complaints, and legal or security questions should always get a person.
  4. Treat "wrong person" as a lead. Thank them, and only reach out to the suggested colleague if that's appropriate and permitted.
  5. Re-run the dry run after changes. Rule edits are easy to make and easy to get wrong.

Illustrative example: A reply says "Not right now, maybe after the new year." A sensible rule classifies it as a timing objection, drafts a short acknowledgment that doesn't push for a meeting, and leaves the draft for review. The reviewer approves it and sets a reminder to follow up later instead of letting the original sequence continue.

What It Doesn't Do

  • It doesn't guarantee a correct classification for every reply
  • It doesn't know facts about your business that you haven't given it
  • It doesn't replace the person accountable for a conversation
  • It isn't a substitute for your own compliance review of what you send

Used this way, response handling clears the routine replies and leaves your team more time for the ones that need a person. For more on designing the handoff, see the reply handoff gap.

Response handling is configured in the Automation step of each campaign. Learn more about the product or see pricing.

Ready to streamline your
client communication?

See how AngelSend's AI-powered platform can help you manage client relationships more effectively.

Starter pack from $3.99 · CAN-SPAM compliant · Reputation protected