<img height="1" width="1" style="display:none;" alt="" src="https://dc.ads.linkedin.com/collect/?pid=332593&amp;fmt=gif">

How to set up chatbots in HubSpot and use AI agents

Renan Andrade
Renan Andrade

Published in: Feb 18, 2025

Updated on: Nov 9, 2026

How to set up chatbots in HubSpot step by step
14:41

Share:

Share on LinkedIn Share on Facebook Share on WhatsApp
Quick answers

About chatbots

Where do chatbots live inside the platform?

Under More > Service > Chatflows, or straight through Service > Chatflows. The Create chatflow button opens the builder, and the Website option publishes the flow on your site pages.

How many bot types are in the library?

Seven: Start from scratch, Concierge bot, Qualify leads bot, Meetings bot, Tickets bot, Knowledge base & support bot and Offline bot. Some of them require a Professional or Enterprise plan.

Is the rule-based chatbot still available?

Yes. The rule-based bot is still documented and available, and the customer agent powered by Breeze sits beside it as an additional path, not a replacement.

Do I need to code to build the flow?

No. The visual editor chains questions, branches and actions as blocks. Technical knowledge only comes in when you want to connect the conversation to an external system.

What you will learn in this article?

This read covers the full conversation setup inside the platform and what changed in HubSpot's automation layer.

  • Where to find chatflows in the current menu and which button starts a new bot.
  • The seven bot types available in the library and the plan requirement attached to each one.
  • How to pick the channel for your bot using your own audience data instead of market averages.
  • The five-step setup, from account access to the final privacy settings.
  • The difference between the rule-based bot and AI agents, including the customer agent powered by Breeze.
  • What Agent Hub is and which agents show up inside it.
  • How to use a HubSpot chatbot for lead qualification, meeting booking and support.
  • Which setup decisions quietly ruin the experience for the person on the other side.
🎯 by the end, you will have the conversation flow designed, the channel chosen from your own data and a clear view of when an agent beats a rule-based bot.
⏱️ Reading time: 13 min
📊 Intermediate
🏢 marketing, sales and service teams already running on HubSpot that want to move the chat out of generic mode.

Chatbots and virtual assistants remain essential for companies that want fast, personalized customer service. Configured well, they change the customer experience and give the team back time for work that needs human judgment.

HubSpot keeps that setup in one place, with a visual editor, a direct link to the contact database and routing rules. The outcome depends less on the tool and more on how the conversation is designed.

A new layer joined the picture. Alongside the bot that follows rules, the platform now offers agents that read content and database records to answer. It helps to understand what changed across the platform's AI tooling before choosing a path.

 

What are chatbots and how do they work?

Chatbots are virtual assistants that interact with users automatically, either through fixed rules or with support from AI (Artificial Intelligence). They answer questions, offer quick solutions and hand the request to a person when the case falls outside the planned script.

Chatbots in 3D: a flow board with connected nodes, chat bubbles and a visored virtual assistant linked to a contact panelCaption: Chatbots handle the first layer of the conversation, and what they record stays in the CRM for whoever takes over.

Beyond saving team time, chatbots support 24/7 service, meaning 24 hours a day on every day of the week. The company stops losing the first reply outside business hours and frees people for more strategic questions.

Inside marketing strategies, they handle lead capture, contact qualification and continuity with people who have not decided yet. Every interaction becomes a record, and every record becomes context for the next approach.

The effect also shows up in broader digital strategies. When the bot asks what someone is looking for, it returns an intent signal that no traffic report delivers on its own.

What are the main features of the HubSpot chatbot?

HubSpot offers a solid way to build and manage automated conversations inside the account itself. The setup is guided, and the core features remain the same:

  • Visual editor: a simple interface for building conversation flows as chained blocks.
  • CRM integration: every interaction is stored on the contact record and gives teams context.
  • Personalization options: greeting, messages, avatar and tone adjusted for each type of audience.
  • Advanced automation: complex interactions are redirected to human agents based on predefined rules.

That last point separates a useful bot from an annoying one. The escalation rule needs to be generous, because insisting on automating something that already became an exception costs more than passing the conversation along.

Teams that want to sharpen how the questions are written will find practical fixes in making the flow sound less robotic, especially in error messages and sign-offs.

How do you choose the communication channel from your audience data?

Every company works with its own mix of channels. Deciding where the bot belongs is not solved by market averages, but by the recorded behavior of your own audience.

The information you need already sits in your account. In the HubSpot CRM (Customer Relationship Management), original source shows how the person arrived, contact properties show what they answered, and conversation reports show where the exchange actually happens.

Compare these readings before deciding: the pages with the most sessions, the channel of the first interaction, and the channel where conversations tend to close. When all three point somewhere different, the bot takes the entry point and the human team takes the closing point.

Audience segment changes the answer. A young audience, roughly between the ages of 18 and 35, rarely concentrates questions on community platforms, so setup effort goes to the social networks and the site chat that group actually uses.

Running that diagnosis first avoids the classic mistake of scattering bots everywhere. Information fragmented across many channels is worth less than complete information in a few, and the HubSpot CRM is exactly where those ends come together.

Video: in this BrainTalks session, recorded in Portuguese, the 4RevOps team shows how the contact database supports channel and automation decisions.

To get more out of the database before writing the first question in the flow, it pays to use CRM records to personalize the conversation with properties your team already fills in.

What are the seven bot types available in HubSpot?

After clicking Create chatflow and choosing Website, the platform opens a library with seven starting points. Some are ready-made templates you edit, and one is the blank canvas.

Bot type

What it does

Plan requirement

Start from scratch

Blank flow, with every question, branch and action built by you

Available in the standard builder

Concierge bot

Greets arrivals, identifies the topic and routes the conversation to the right team

Service Hub Professional or Enterprise

Qualify leads bot

Asks qualification questions and writes the answers into contact properties

Available in the standard builder

Meetings bot

Sends the visitor straight to a rep's calendar and confirms the meeting

Available in the standard builder

Tickets bot

Opens a support ticket from the conversation itself, with the history attached

Available in the standard builder

Knowledge base & support bot

Surfaces knowledge base articles and offers human help when the article falls short

Service Hub Professional or Enterprise

Offline bot

Replies outside service hours and collects contact details for a follow-up

Sales Hub or Service Hub Professional and Enterprise

Table: the seven starting points in the chatflow library, with the plan requirement attached to each one.

The template you pick defines the work that follows. Starting from the Qualify leads bot means inheriting the question structure and only adjusting vocabulary. Starting from scratch means controlling everything, mistakes included.

If the tooling decision is still open, comparing platforms before committing saves migration work later on.

How do you set up chatbots in HubSpot step by step?

The setup is a guided process, and the sequence below runs from first access to privacy rules. Follow the five steps in order so you do not have to backtrack.

Step 1: account access

Log in and go to More > Service > Chatflows. In accounts where the service menu is already visible, the direct Service > Chatflows path lands in the same place. Click **Create chatflow** and choose the **Website** option.

If that screen does not appear, the likely cause is an inbox that has not been connected yet. The platform walks you through the connection, and the knowledge base carries a dedicated tutorial on the shared inbox.

Both the bot and live chat only show on pages that load the HubSpot tracking code. Confirm the code is installed before moving on, because without it the published flow will not render.

Step 2: organizing the flow

Define where the flow will run. With the Website option selected, you can browse the templates already in the account or start from scratch. Templates accept full customization, so starting from one limits nothing.

Next, select the inbox that will receive the messages and the language of the conversation. That inbox determines who gets notified when the chat escalates to a person.

Step 3: selecting actions

Build the conversation by clicking the [+] button and browsing the available actions. The ones used in almost every flow are:

  • Send a simple message;
  • Question;
  • Branch;
  • Set property value;
  • Send a ticket;
  • Send to team members;
  • Book a meeting.

Explore all the options and confirm that each block pushes the conversation toward your goal. The "go to action" feature opens the details of each stage, and clicking the action box lets you review the current setup without dismantling the flow.

Pay extra attention to the set property value action. That is what turns a visitor answer into usable data instead of leaving the information buried in the conversation history.

Step 4: pages defined

Open the targeting tab and define which pages show the flow. You can release it site wide, restrict it to a specific page, or exclude addresses where chat gets in the way, such as checkout and logged-in areas.

Then select the visitor behavior conditions that decide who sees the bot. You can display it only to contacts on a specific list, or only to people browsing from a certain device.

In the display tab, set the title, the avatar and how the message opens:

  • A pop-up showing a preview of the text;
  • The bot icon only;
  • Automatic opening as soon as the visitor lands on the page.

You can also trigger the bot when the user tries to leave the page or after a set amount of scrolling. Adjust the accent color, the placement and, above all, the behavior on small screens, where chat takes over the view easily.

Step 5: final settings

Open the remaining options and handle the details that usually get skipped:

  • Session timeout and automatic restart at the interval you choose;
  • Delay between messages, so the conversation does not read like a broadcast;
  • A fallback message for when the flow hits an error;
  • The language this bot is being configured in.

Set availability according to team hours, or keep the bot running full time. Finally, handle privacy and consent by enabling cookie collection and data processing in line with the privacy rules that apply to your market.

What changes between the rule-based chatbot and AI agents?

The rule-based bot follows the path you designed. It asks, branches and executes what the map says, with no improvisation. That predictability is an advantage in qualification, meeting booking and ticket creation.

You can also create a customer agent powered by Breeze, which uses CRM content and data to answer. The rule-based bot stays available beside it, with no deprecation notice.

The practical difference sits in where the answer comes from. The rule-based bot replies with what the flow says. AI agents assemble the reply from the material you connected, which covers questions nobody anticipated in the original design.

Video: practical reasons to invest in AI-powered conversations for acquisition and retention, recorded in Portuguese.

For commercial operations, the most direct case is the AI SDR agent. SDR (Sales Development Representative) is the pre-sales role that makes first contact and qualifies the opportunity.

An AI SDR agent takes on part of that work at scale. Look at when the agent takes over voice and messaging prospecting before splitting tasks between bot, agent and team.

In an inbound marketing strategy, AI agents cover the open questions a visitor brings before considering a purchase. The rule-based bot stays better at structured capture, when you know exactly which fields you need filled.

In an inbound sales strategy, the logic flips: the agent pulls contact history and answers with context, while the rule-based bot makes sure the meeting lands on the right calendar. Running both usually beats picking a side.

Anyone who wants the difference between a rule-based bot and an agent applied to student recruitment will find the full breakdown with flow examples.

What is Agent Hub and which agents appear in it?

Agent Hub gathers every agent and agentic workflow in the account into a single panel. That is where the team sees what exists, activates it, configures it and monitors each result.

The agents available are:

  • AEO (Answer Engine Optimization), which requires Marketing Hub Professional or Enterprise;
  • Data agent, focused on handling the database;
  • Prospecting agent, focused on outbound research and outreach;
  • Deal progression, focused on moving deals forward;
  • Customer agent, the service agent powered by Breeze.

Availability varies by product. Marketing, Sales, Service, Data and Content Hub appear on Starter, Professional or Enterprise plans. Smart CRM and Revenue Hub appear on Professional or Enterprise.

Usage consumes HubSpot Credits, and some agents, along with agentic workflows, require Professional or Enterprise. Checking that before designing the operation avoids promising the team automation the current plan does not deliver.

The practical reading is simple: the panel shows what is running and what is idle. Active agents without monitoring turn into cost without return, and Agent Hub exists precisely to close that loop.

How do you use a HubSpot chatbot to improve customer service?

Once the flow is published, the bot takes on concrete jobs inside marketing strategies and automated service:

  • Lead qualification: identifies whether the visitor is ready for a sales approach and writes the stage onto the contact record;
  • Meeting booking: sends the person straight to a rep's calendar, with no back and forth to find a slot;
  • Frequently asked questions: answers repeat questions and reveals, through volume, what needs fixing on the site or in the product;
  • Promotional campaigns: presents offers inside personalized interactions, based on what the contact already shared;
  • Technical support: resolves simple cases and escalates complex ones to the specialist team with the history attached.

The gain shows up in the time between a question and the first useful answer. It is worth measuring the effect on service response time before and after the flow goes live.

Handing the conversation to a person is the most fragile point of the whole operation. Defining queues, owners and deadlines helps organize service once the bot passes the conversation on without leaving anyone waiting.

That arrangement supports digital strategies that depend on fast replies. In an inbound marketing strategy, the chat holds the question that would otherwise close the tab. In an inbound sales strategy, it hands over an opportunity already qualified.

Frequently asked questions about chatbots

More > Service > Chatflows, or Service > Chatflows directly. Then the Create chatflow button and the Website option to publish the flow on site pages.

No. The rule-based bot walkthrough still stands, and the customer agent powered by Breeze is an additional option inside the same account.

Concierge bot and Knowledge base & support bot require Service Hub Professional or Enterprise. Offline bot requires Sales Hub or Service Hub Professional and Enterprise.

Yes. The common arrangement keeps the rule-based bot on structured capture and the agent on open questions, with escalation to the team when the case needs human judgment.

HubSpot Credits in the account. Some agents and the agentic workflows also require Professional or Enterprise, and the AEO agent requires Marketing Hub Professional or Enterprise.

Look at contact source and the channel of conversations already recorded in your HubSpot CRM. The channel that concentrates your audience's first interaction is the starting point.

No. A flow published for the website only renders on pages that load the HubSpot tracking code, so the installation has to be confirmed before testing.

How do you make the chatbot part of your operation?

Setting up chatbots in HubSpot stopped being an isolated service project. The rule-based flow delivers predictability, AI agents cover what nobody anticipated, and Agent Hub brings both worlds into the same panel.

The next move is connecting that conversation to the rest of the operation: properties filled in, service queues defined and reports reviewed on a regular cadence. For help designing that arrangement, talk to the mkt4edu team.

Discover the best platforms to create chatbots!

Join us!

Did you like this content? Share it!

Technologies we use

The world changes all the time and technology is no different! Here at Mkt4Edu, technology is in our DNA, we work with many different softwares to make the whole process of automation and artificial intelligence work more efficiently and achieve more results.

Here, new softwares are tested all the time. Modern tools and new functionalities are tested all the time, there were already more than 200 tests so you can have the best result in your institution.


From customer acquisition to retention: Mkt4edu can make the difference in your marketing operation.

captacao_leads

Increase your leads’ capture

retencao_clientes

Improve your customers’ retention

reducao_custos

Save conversion costs