# Lindy AI: How to Automate Your Inbound Email Replies

*How-to — 2026-10-24 — by Mahmoud Zalt*

Set up Lindy to handle inbound email: labelling rules, drafted replies, urgent-message alerts, follow-up bumps, and what it never sends on its own.

**Short answer.** **Lindy handles inbound email through built-in routines rather than a workflow you build: labelling first, then drafting.** You create at least one label with instructions for when it applies, turn the routine on, then enable email drafting so replies land in your drafts folder. Their documentation is clear that drafts are not sent automatically, so you keep the final click.

**About this guide.** This guide is based on publicly available information and Lindy's own documentation, checked on 17 August 2026. Sistava is not affiliated with Lindy. Interfaces and pricing change often, so confirm the current steps in their documentation before you rely on them.

There are two different jobs hiding inside handling your email, and mixing them up is why most attempts fail. The first is deciding what each message is. The second is deciding what to say back. Lindy separates them into two routines that run in order, and understanding that order makes the whole feature behave predictably.

Labelling comes first, and drafting follows it. Their documentation describes drafting as replying to the emails that need responses after triage has run. So if drafts look wrong, the cause is often a labelling rule rather than the drafting prompt. Fix the sorting before you rewrite the tone instructions.

## How do you set up email labelling in Lindy?

You create labels in a drag and drop interface split into two sections. One is described as moving messages out of your inbox, which applies the label and archives the message. The other keeps them in the inbox, applying the label while leaving the message visible. Which section a label sits in decides whether it clears your inbox or just organises it.

Each label carries four things. A name, though Gmail's reserved names such as Inbox and Sent are rejected. A colour. Instructions describing when to apply it, with their own example being any email from an investor. And an optional toggle to remove the label once you reply, which keeps a to-respond pile honest.

You do not start from nothing. The documentation lists six default labels: To Respond, FYI, Newsletters, Comments, Notifications and Invoices. All of them can be edited or removed, so the fastest setup is to keep the ones that match how you already think and rewrite the instructions on the rest.

1. **Connect the inbox** — Labelling needs a connected Gmail or Outlook account. Their documentation also states it requires a plan that includes the Assistant features, and that an admin can disable it for a workspace.
2. **Create at least one label** — The routine cannot be enabled until one label exists. Start with the default set and adjust rather than building from scratch.
3. **Write the instruction, not the keyword** — The instructions field describes when a label applies, in plain language. Something like any email from an investor is closer to how it works than a list of search terms.
4. **Decide which labels archive** — Drop a label into the section that moves messages out of the inbox only if you are comfortable not seeing them. Newsletters usually belong there. Anything from a customer usually does not.
5. **Turn the routine on** — Enable it with the toggle on the routine card or the switch in its drawer. It then runs automatically on incoming mail.
6. **Watch a day of real mail** — Check what got labelled and what got archived before you enable drafting on top of it. A labelling mistake becomes a drafting mistake.

## How do drafted replies work, and will they send themselves?

They will not send themselves. Their documentation states directly that Lindy does not send email automatically and that nothing goes out without your approval. Drafts are written in your style, placed in your drafts folder, and wait there for you to read, edit and send. That is a deliberate design choice and a reassuring one.

The settings sit on the Routines page, in an Email drafting card. Their documentation names four things you control there: drafting instructions, where you set tone, formality, length and the conditions under which it should draft or skip a reply; your signature, a custom sign-off that supports Markdown formatting; learned instructions, which capture patterns from your actual emails and need saving manually after you edit them; and a toggle for the Sent with Lindy footer, whose availability depends on rollout.

Spend your time on the drafting instructions, and especially on the skip conditions. Telling it when not to draft is more valuable than telling it how to write, because a missing draft costs you nothing and a confident draft on a delicate thread costs you your attention at the worst moment.

**If that felt like a lot of configuration.** Tuning label instructions, then tone rules, then skip conditions, is exactly the setup work we removed at Sistava. You hire an AI Employee, tell it in plain English which mail it owns and which it must always bring to you, and it works from that brief. It has the inbox connected already and asks you when a message does not fit what you described.

## How do you get told about the urgent ones?

Through a separate routine called Email alerting, described on its card as alerting you to time-sensitive emails. You enable it from the Personal tab of the Routines page and write alert instructions describing what counts as urgent. Their own examples include alerting only for meetings within the next half hour, for mail from a specific important client, or for replies on a named thread.

The alert can reach you by text message, Slack direct message, Lindy chat, or your default proactive channel. This is the routine that makes the rest safe to leave alone, because it means an archived inbox does not also mean a missed emergency. Set it up before you get comfortable with labelling, not after.

## What about the emails you sent and never heard back on?

That is a fourth routine, Follow-up bumps, described as drafting a bump when sent emails get no reply. It watches your sent mail and, when a message goes unanswered past your chosen window, writes a follow-up into your drafts. Like drafted replies, it does not send on its own.

The timing setting is called draft follow-ups after, and the documented choices are never, one day, two, three, five or seven days. Their documentation also notes this routine runs independently of email drafting, so you can use one without the other. If you only ever adopt one of the four routines, this is a reasonable candidate, because nobody has ever regretted a chased proposal.

**Limits the setup screens do not spell out.** These are built-in routines, and their documentation says so pointedly: labelling does not expose the extras a custom routine would, meaning no custom triggers, no schedules and no run history. It runs on all incoming mail rather than per message, so there is no per-email switch. Settings are also per inbox, selected by account chip, so a second Gmail does not inherit the first one's rules and has to be configured again. And connected inboxes are capped by plan, starting at two on the entry paid tier on the pricing page we read on 17 August 2026.

## When should you build a workflow instead?

When the mail is not yours. The routines above are built for a personal inbox. For a shared address that answers customer questions, the workflow builder is the better fit, because it gives you the Gmail triggers directly, including Email Received with filters, New Email Matching Search and New Labeled Email, plus actions for Send Reply and Draft Email.

An independent walkthrough of that route adds two useful details. You can attach a knowledge base of documents, pages and structured content such as pricing, so answers come from your material rather than from guesswork. And a Save as Draft option decides whether replies go out or wait for you, which is the same safety switch the built-in routine hardcodes.

Whichever route you take, the same rule applies: test against real messages before you trust it, including the awkward ones it should refuse. That is also the bet behind Sistava. An AI Employee that owns your inbox is briefed like a person, told which mail it may answer and which it must always escalate, and it checks with you rather than assuming. Different shape, same principle, which is that the human keeps the decisions that matter.

## FAQ

### Will Lindy send emails without asking you?

Not through the email drafting routine. Their documentation states that Lindy does not send emails automatically, that drafts land in your drafts folder, and that nothing goes out without your approval. If you build a custom workflow instead, that workflow can include a send action, so the safety depends on how you configure it.

### What are Lindy's default email labels?

Their documentation lists six: To Respond, FYI, Newsletters, Comments, Notifications and Invoices. All of them can be edited or removed. You must have at least one label before the labelling routine can be enabled, so the defaults double as the fastest way to get started.

### Can Lindy archive email automatically?

Yes, depending on where you place the label. The labelling interface has one section that applies a label and moves the message out of the inbox, and another that applies the label while keeping the message visible. Newsletters are the usual candidate for the archiving section. Anything from a customer usually is not.

### How do you make Lindy write in your voice?

Two settings do the work. Drafting instructions is where you specify tone, formality and length, and learned instructions holds writing patterns captured from your own emails, which you can edit and must then save manually. The documentation also notes it learns from the edits you make to its drafts over time.

### Can you use different rules for a work inbox and a personal one?

For the inbox routines, yes: their documentation describes connecting multiple Gmail or Outlook accounts with settings kept separate per inbox, selected by account chip. Note that the integrations documentation states a different limit, that you cannot yet set one guardrail rule for a work Gmail and another for a personal Gmail, so the two areas behave differently.

### How long does Lindy wait before drafting a follow-up?

You choose. The setting is called draft follow-ups after, and the documented options are never, one day, two days, three days, five days or seven days. The routine only drafts the bump, it does not send it, and it runs independently of the email drafting routine.

## Sources

- [Lindy documentation, ](https://docs.lindy.ai/features/inbox-management/email-triage)Email Labeling
- [Lindy documentation, ](https://docs.lindy.ai/features/inbox-management/email-drafting)Email Drafting
- [Lindy documentation, ](https://docs.lindy.ai/features/inbox-management/email-alerting)Email Alerting
- [Lindy documentation, ](https://docs.lindy.ai/features/inbox-management/follow-up-bumps)Follow-up Bumps
- [Lindy documentation, ](https://docs.lindy.ai/skills/popular-integrations/gmail)Gmail integration
- [Lindy documentation, ](https://docs.lindy.ai/integrations/overview)Integrations
- [Lindy documentation, ](https://docs.lindy.ai/pricing)Pricing, read 17 August 2026
- [FlowHunt, ](https://www.flowhunt.io/blog/how-to-use-lindy-ai-beginner-guide/)How to Use Lindy AI, a beginner's guide

**Tags:** lindy-ai, email-automation, inbox-management, ai-email-assistant, ai-assistant-setup