Help Scout

Set up an AI agent for Help Scout that drafts or sends replies in your shared mailbox, trained on your Docs site and past conversations like your team.

eesel's AI agent for Help Scout works inside your Help Scout mailbox. It drafts or sends replies, adds notes and tags, and learns from your Docs site and past conversations.

Help Scout keeps everything in a shared mailbox rather than a ticket queue. Your AI teammate works the same way, picking up conversations beside your team, inside Help Scout. It reads your Docs site and your past conversations first, so it knows how you answer before it touches anything. Then it handles conversations the way you tell it to. It can park a draft in the reply box for a person to check. It can answer the routine ones itself. It can tag and update the rest.

In this example the agent answered as an internal note and named the Docs page the answer came from.

Your agent reads the conversation, searches your Docs site, and writes up what it found.

Set up eesel in Help Scout

1

Ask your agent to connect Help Scout

Open the chat in your dashboard and ask it to connect Help Scout. There is one way to connect. The agent points you at your Integrations page, because the authorizing happens there. Prefer to click? Integrations in the left sidebar takes you to the same place.

The agent links you to the Integrations page and waits there.

2

Open Help Scout under Integrations

The Integrations page lists everything you can connect. Type help in the Search integrations box to find the Help Scout tile.

Searching for help brings up the Help Scout tile.

Click the Help Scout tile. Its page opens. Click Connect at the top right.

The Help Scout page, with the Connect button at the top right.

3

Authorize it in Help Scout

eesel uses a service called Pipedream to handle the authorization with Help Scout. A window explains that and asks you to continue. You do not need a Pipedream account of your own. Click Continue.

Pipedream handles the authorization between eesel and Help Scout.

Help Scout then asks you to authorize Pipedream. Click Authorize. If you are not signed in, Help Scout asks you to sign in first.

Help Scout asks you to authorize the connection.

The Help Scout screen says everything in the Help Scout API becomes available. The connection acts as the Help Scout user you signed in as. So your agent sees what that user sees, and nothing more. Sign in as a user whose access matches what you want your agent to have.

This approval is the one step a person has to do in the browser. If an AI agent is doing the setup for you, it hands you the link and waits.

4

Confirm the connection

Pipedream confirms the account is connected. Click Continue to close the window.

The connection is done.

The Help Scout page behind it now shows Connected, and your conversations start syncing.

5

Check what it can read

Integrations > Help Scout in the left sidebar shows the connection and its sources. Each source has its own toggle, and shows a document count once its sync is done. From a terminal, eesel status lists the integration as connected with its source counts.

Conversations has synced 324 documents. Help Center is still at zero.

Conversations syncs on its own. It indexes every mailbox by default, and you can restrict it to specific mailboxes.

Help Center is your public Help Scout Docs site. It needs the site's URL, for example https://docs.example.com. Click the gear icon on the Help Center source and enter it there. Nothing is crawled until you do. That is why Help Center shows 0 docs above.

Scroll further down the same page for the Actions section. That is where you choose what your agent may do.

6

Create your first automation

Connecting alone does not start your agent. It needs an automation first. Tell it what to handle and it creates an automation. An automation is a Help Scout trigger plus instructions. In the instructions you describe which conversations it should handle and what to do with them. It skips the rest.

Ask in chat for the automation you want, and the agent builds it.

It shows up on your Automations page, switched on and ready. From a terminal, eesel automations lists it as on.

The new automation on your Automations page, switched on.

Open the automation to see or change what it does. This one runs on A customer opens a new conversation. Its instructions tell the agent to read the conversation and create a draft reply, without sending it. The trigger event is fixed when the automation is created. So a new automation is required for each different trigger event.

The automation editor, where you change the instructions it follows.

Send reply always sends. It has no draft mode, and a sent reply cannot be recalled. To keep a person in the loop, use Create draft reply instead. It saves the reply unsent in the conversation's reply box, and a teammate reviews, edits and sends it from Help Scout.

7

Check automation work in Help Scout

Open Help Scout and look at the next conversation that comes in to see what the automation did there. In this example the agent left its reply waiting in the conversation's reply box. A teammate reviews it and sends it.

Help Scout attributes the draft to the user you authorized, so the author shows as You. There is no line saying eesel AI left it. The proof is the Draft label on a reply nobody typed.

The customer email, with the drafted reply under it marked Draft.

Scroll down and the whole draft is there, from the greeting to the sign-off, ready to edit and send. An Edit button sits under it in Help Scout.

The full draft reply in the conversation, labelled Draft, with the Edit button below it.

From Claude Code, Cursor or a terminal. The same setup works as commands: eesel login, then eesel integrations connect help_scout (a person approves it in the browser; if the receipt has no link, ask eesel chat for one), eesel integrations download start to copy your sources in, and eesel automations to see and switch on the automations. Trigger keys for this page are in the Triggers table below. Full command reference: The eesel CLI. Step-by-step: Set up from Claude Code or the terminal. From an MCP client: Claude Code and MCP clients.

Sources, triggers and actions

Sources

Sources are what it learns from. Toggle each one on or off from the Help Scout page under Integrations:

Source	Description
Conversations	Your past conversations. Every mailbox by default, restrictable to specific mailboxes
Help Center	Your public Help Scout Docs site, crawled from the URL you enter on the source

Every other integration you connect is also a source. So a Help Scout reply can draw on Confluence, Notion, your website, and your files at once. See Integrations.

Triggers

Triggers are the Help Scout events an automation can run on:

Trigger	Fires when	Best for	Key
A customer opens a new conversation	A new conversation is created	Drafting a first reply, triage and routing	help_scout_new_conversation
A customer sends a message	A customer replies on an existing conversation	Following a conversation through to the end	help_scout_customer_reply

Actions

Actions are what it can do. You find them in the Actions section of the Help Scout page under Integrations. Help Scout's actions come from the live connection rather than a fixed list in eesel. So you may see more than the ones below:

	Action	What it does
Respond	Create draft reply	Saves a reply unsent in the conversation's reply box, for a teammate to review, edit, and send. Nothing reaches the customer. Most teams start here
	Send reply	Sends a reply to the customer. This always sends, and a sent reply cannot be recalled
	Add note	Adds an internal note only your team sees
Manage	Update conversation	Updates a conversation's tags, status, or assignment
	Create customer	Creates a new customer record in Help Scout
Look up	List tags	Lists the tags in your Help Scout account

You stay in control the whole way. Adjust your AI agent's access to knowledge and its action permissions to your liking. See Actions and Approvals.

Limitations

• Live Beacon chat is not supported. Beacon messages sent in email mode arrive as normal conversations, and those work.

• Drafts show the authorized user as the author. Help Scout has no way to label a draft as written by eesel AI.

• One trigger per automation. The trigger event is fixed when the automation is created, so each trigger needs its own automation.

• Help Center needs its URL. Nothing on your Docs site is crawled until you enter the URL on the Help Center source.

• Sent replies cannot be recalled. To keep a person in the loop, use Create draft reply instead of Send reply.

• No sender blocklist. The first check cannot see the sender's email address. To skip a sender or domain, tag their conversations with a Help Scout Workflow, for example no-ai, and tell the automation to skip that tag.

• Mailboxes are named by ID. To keep an automation to one mailbox, name the mailbox ID (the number in its Help Scout URL) in the instructions.

Troubleshooting

Connection didn't complete?

Complete every prompt in the Pipedream window. Closing it partway leaves the connection incomplete. Start again from the chat, or from Integrations in the left sidebar.

Agent not replying to conversations?

Check the automation is enabled on the agent's Automations page. Check the automation's instructions cover the conversations you expect. Check the connection is still authorized, and that Create draft reply or Send reply is turned on.

Docs articles not showing up?

Nothing is crawled until you enter your Docs site URL on the Help Center source. Confirm the URL is correct and the site is public.

Past conversations not training?

Confirm the connection is authorized, then re-sync the Conversations source. If you restricted indexing to specific mailboxes, check the conversations you expected are in one of them.

Only actions are available, no sources or triggers?

You may have connected Help Scout (API Keys), which is a separate integration marked Actions only. Connect the plain Help Scout integration to get sources and triggers as well.

FAQ

How do I add an AI agent to Help Scout?

Connect Help Scout from the Integrations page in eesel, approve the authorization in Help Scout, then ask your agent in chat to create an automation, for example "draft a reply to every new conversation". Connecting alone does nothing: your agent only acts once you create an automation and enable an action. The authorization runs through Pipedream, a service that handles the connection between eesel and Help Scout, and you do not need a Pipedream account of your own. The steps are in Set up eesel in Help Scout.

How long does it take to set up an AI agent in Help Scout?

About 30 minutes. That was the median time from connecting Help Scout to the agent working on a live conversation, across real customer setups, including the wait for the first customer email. The work is one browser authorization in Help Scout and one automation, which you can ask your agent to create in chat. Your past conversations sync on their own.

Does the Help Scout AI agent send replies or only draft them?

Either, depending on how you set it up. Create draft reply saves the reply unsent in the conversation's reply box and sends the customer nothing, so a teammate reviews, edits and sends it from Help Scout. Send reply puts the message straight in front of the customer and cannot be recalled.

Help Scout attributes a draft to the user you authorized during setup, so your team sees the author as You. Help Scout has no way to label it as eesel AI.

How do I keep a human in the loop with the Help Scout AI agent?

Use Create draft reply instead of Send reply. Your agent saves the reply unsent, and your team reviews, edits and sends it from Help Scout. See Actions and Approvals for choosing what your agent may do.

Can I limit which Help Scout conversations the AI agent handles?

Yes. In the automation's instructions, describe the conversations you want in plain language and what to do with them. It takes account of the conversation's tags, its mailbox, its subject, and whether the message looks automated. To skip a sender or domain, tag those conversations with a Help Scout Workflow and tell the automation to skip the tag, as in Limitations. Skipped conversations are recorded on the agent's Activity page, so you can see what it passed over and why.

Can I train the Help Scout AI agent on past conversations, and where does the data go?

Yes. Open your Help Scout integration and sync the Conversations source. The content trains your own agent inside eesel and goes nowhere else. Your agent also learns from your Help Scout Docs site and anything else you connect, such as Confluence, Notion, Google Drive or your website.

Can I connect Help Scout to Claude or ChatGPT?

Yes, through eesel. Once Help Scout is connected, you can talk to your eesel agent from Claude, ChatGPT or Cursor, and it works with your Help Scout data. See Use eesel from Claude and other assistants, Use eesel from ChatGPT and Claude Code and MCP clients.

What reporting do I get for the Help Scout AI agent?

The Activity page lists every conversation your agent handled, with a link to it in Help Scout, what it searched, each action it took, who approved it, and its reasoning. The Reports page shows volume, triggers, and approvals over time. You can also ask your agent for a written report, or put one on a schedule.

How do I disconnect Help Scout from eesel?

Open Help Scout under Integrations in the left sidebar, then use the ... menu at the top right to remove the connection. Your agent loses both the sources and the actions straight away.

Related pages

• Helpdesk for the job your teammate takes on in a ticket queue

• Automations and Schedules for triggers, instructions and schedules

• Actions and Approvals for choosing what your agent may do

• Integrations overview for how sources, triggers and actions fit together

• Set up from Claude Code or the terminal