Ever feel like you’re trying to manage way too many conversations all at once? In today’s busy world, people really expect quick answers, whether it’s the middle of the workday or late at night. This is where chatbots become incredibly useful, acting like helpful team members who are ready to assist around the clock. And with AI joining the party, these aren’t just simple bots following a script anymore; they can actually understand things better and give more helpful responses.
HubSpot, a platform many businesses already use for keeping track of customers and running marketing campaigns, offers tools to build these conversational helpers. They started with more basic bots that just followed rules, but HubSpot has been adding AI features to make them smarter over time.
This guide is here to help you walk through building a chatbot using HubSpot’s Chatflows tool. We’ll cover everything from getting things set up to figuring out how the conversation should flow, and we’ll point out where HubSpot’s AI features fit in. We’ll also chat a bit about how specialized AI solutions, like eesel AI, can take support automation even further, especially if you’re looking for deeper smarts and more flexibility than what’s built into HubSpot.
What you’ll need to get started
Before you jump into building your chatbot in HubSpot, there are a few things you’ll want to have ready. Think of these as the must-have items for your project:
- A HubSpot account. The specific chatbot features you can use might depend on the plan you have, especially when it comes to more advanced stuff like connecting to your knowledge base or using certain AI features, as the context information mentions.
- The right permissions inside your HubSpot account. Specifically, you’ll need Chatflows permissions under the CRM tab so you can create and change chatflows, according to HubSpot’s own documentation.
- A connected chat channel or inbox set up in HubSpot. This is where your chatbot conversations will happen and where you’ll manage everything.
- If your website isn’t hosted directly on HubSpot, you’ll need to make sure the HubSpot tracking code is installed on the pages where you want the chatbot to show up. If that code isn’t there, the bot won’t appear.
Understanding Hubspot’s chatbot capabilities
HubSpot gives you a few ways to automate conversations, and it helps to know how they work before you start.
The main tool is Chatflows, where you build bots that follow rule-based flows—like a script you map out using clicks and responses. These are great for simple tasks like answering common questions or guiding users to the right page.
HubSpot has also added AI-powered features. Most notably, you can now connect your bot to your Knowledge Base, so it can pull helpful articles instead of just sticking to the script.
In contrast, HubSpot’s AI Assistants (or Copilots) are for internal use. These help your team draft emails, summarize content, or find info but they don’t power your customer-facing chatbot.
If you’re looking for smarter automation, tools like eesel AI go further. They can train on support tickets, internal docs, and over 100 sources, not just your HubSpot Knowledge Base, giving the bot more context and accuracy.
Knowledge source comparison:
- HubSpot AI: Pulls only from your HubSpot KB articles.
- eesel AI: Pulls from help desks (like Zendesk, Intercom, Freshdesk), tickets, docs, FAQs, and wikis.
Step-by-step guide to building your Hubspot AI chatbot
Ready to put together your conversational helper? Let’s walk through the steps in HubSpot, one by one.
Step 1: Find the chatflows tool
Your journey starts inside your HubSpot account. Once you’re logged in, head over to the main menu.
You’ll want to find the “Conversations” section. Click on that, and then pick “Chatflows” from the list that drops down. This is your main spot for creating and managing all your chatbot and live chat setups in HubSpot.
Step 2: Make a new chatflow
Now that you’re in the Chatflows area, you’ll see a button to create a new chatflow. Go ahead and click it.
HubSpot will ask you what kind of chatflow you want to make. For a chatbot on your website, choose “Website.” (You can also make bots for Facebook Messenger, but we’re focusing on websites here). On the next screen, you’ll pick the type of chat experience. Select “Rules-Based Chatbot.” Even though it says “Rules-Based,” this is the tool that lets you build a bot and include HubSpot’s AI features like linking to your Knowledge Base.
You’ll choose between:
- Website: For embedding the chatbot on your website pages.
- Facebook Messenger: For connecting the chatbot to your Facebook page.
And then select the chat experience:
- Live chat: For connecting visitors directly to a human agent.
- Rules-Based Chatbot: For building an automated conversation flow (this is the one you need for a chatbot).
Step 3: Set up your chatbot’s main actions
After you’ve picked the type, you’ll land in the chatbot builder area. First, choose the inbox or help desk you want to connect this chatflow to and set the language.
Next, you can either pick a template or start from scratch. Either way, you’ll need to set up the first message that greets visitors.
Now for the fun part: adding actions. Click the “+” icon in the builder to add steps to the conversation. You can add things like:
- Asking questions to get information from the visitor.
- Setting details about the contact based on their answers.
- Sending the chat to a specific team member or team.
- Adding a “Knowledge base lookup” action.
This “Knowledge base lookup” is where HubSpot’s AI feature within the builder really helps, letting the bot search your connected HubSpot Knowledge Base and share relevant articles or summaries.
While it’s helpful that HubSpot can pull from its own Knowledge Base, it’s worth mentioning that specialized AI solutions like eesel AI can learn from a much wider variety of sources than just the HubSpot KB. This includes things like past support tickets, internal company documents, and outside wikis, which can lead to even more complete and understanding answers.
Step 4: Design how your conversation flows
The builder gives you a visual way to map out the conversation path. You’ll connect the actions you added in the last step to create a logical flow.
To make your conversations dynamic, use “if/then branches”. These let the bot go down different paths depending on what a user says, a detail about the contact, or other conditions. For example, if someone says their issue is “Billing,” the flow can branch off to ask billing-specific questions or show relevant Knowledge Base articles. Just keep in mind that advanced branching might have limits on lower HubSpot plans.
It’s a good idea to design paths that make sense and try to guess common questions and situations. Map out how the chat should go for different reasons people might contact you, and make sure there are clear ways for users to get the help they need.
Building detailed, smart flows that can handle subtle differences, understand lots of different reasons for contacting you, and connect with many outside systems is where more advanced platforms like eesel AI really shine. They offer more flexibility and power than just standard rule-based logic or simple AI lookups within one platform.
Step 5: Customize how it looks and its settings
Once you’ve mapped out how your conversation should go, head over to the “Display” tab. Here, you can change how the chatbot looks on your website.
You can customize:
- The chat picture and the title at the top of the chat window.
- When the bot shows up (decide if the welcome message pops open right away or if just the little chat icon appears).
- Rules for when the bot appears, like after someone has been on the page for a certain amount of time or when they look like they’re about to leave.
Next, check out the “Options” tab. Here you can set things like:
- How long the bot pauses before typing.
- How long a chat session lasts before it resets.
- The general message users see if something goes wrong.
- The bot’s language.
- If it should be available based on your team’s working hours (Context 2).
- Data privacy settings, including asking for cookie consent and permission to use data (Context 2).
Step 6: Set rules for where it appears
Now, you need to tell HubSpot exactly where you want your chatbot to show up. Go to the “Target” tab.
You can set rules based on:
- Specific web page addresses.
- Pages that include certain words.
- Information you already know about your visitors, like details you have in their contact profile.
Use inclusion rules to say where the bot should be and exclusion rules to stop it from showing up on certain pages, like your privacy policy or checkout page. This lets you make the chatbot experience just right for different parts of your website or different groups of visitors.
Step 7: Test it out and turn it on
Before you let your chatbot loose on the world, it’s really important to test it completely. Use the “Preview” function in the builder to go through the conversation flow just like a visitor would. Try out different things people might type and different situations to make sure the bot answers the way you expect and that the branches work correctly.
Once you feel good about it, go back to the main chatflows list and click the little toggle switch next to your chatbot’s name to turn it on. It will then start appearing live on the pages you picked in your targeting rules.
While HubSpot’s preview is helpful for checking the flow, it doesn’t quite feel like real conversations. In fact, tools like eesel AI offer better testing options. They let you simulate how the bot would respond to actual past support tickets or specific tricky situations before you make it live. This gives you more confidence that it’s accurate and helps you guess how well it will work.
Tips and best practices for your Hubspot AI chatbot
Building the chatbot is just the start. To make it work really well, especially when you’re using its AI features, here are some tips and good practices to keep in mind:
- Define its purpose: Be clear on what your chatbot is solving, support, lead gen, bookings, or something else.
- Know your audience: Match the bot’s tone and flow to how your users talk and what they usually ask.
- Organize your content: Keep your HubSpot Knowledge Base accurate and updated. Or use eesel AI to pull from wider sources like past tickets or internal docs.
- Keep flows logical: Build clear paths for common questions and always allow an easy option to reach a human.
- Review and refine: Use reports to spot where the bot struggles. Tools like eesel AI’s Knowledge Gap Analysis can help you improve its answers.
- Enable smooth handoffs: Set up a clear handoff process, so agents can jump in with full context.
- Involve the team: Coordinate with support, marketing, and devs to keep the chatbot useful and on-brand.
Overcoming Hubspot AI chatbot limitations with eesel AI
While HubSpot gives you a good starting point for building chatbots and includes some AI features, there can be limitations, especially if you want really advanced support automation.
Here are some key areas where HubSpot’s built-in AI features might fall short compared to specialized solutions like eesel AI:
- Where the bot can learn from: HubSpot’s built-in AI mostly uses its own HubSpot Knowledge Base. If your important support information is in other places like past tickets, internal documents, or outside wikis, you might find that HubSpot’s AI has trouble getting to and using that information well. On the other hand, eesel AI is built to learn from over 100 different sources, including past tickets, Google Docs, Confluence, and lots more. This gives its AI agent a much wider and deeper understanding.
- Basic customization: While HubSpot offers some preset ways the bot can sound, you might find it hard to really fine-tune the bot’s personality or make specific answers match your brand’s exact voice or handle tricky situations. eesel AI lets you control the tone in detail and lets you give specific instructions on how the bot should respond in unique cases.
- Testing also has limits: HubSpot offers a preview feature, but it doesn’t let you really test how the bot would handle actual past tickets or complicated situations before it goes live. eesel AI provides advanced testing features, letting you see how the bot performs on historical data and even roll it out to just a few agents first. This helps you be more sure it’s accurate and predict how well it will work.
- Limited ability to take action: Plus, HubSpot’s built-in actions can sometimes lack deep ability to take action. While it can set properties or create tickets, doing more complex things like getting order details from Shopify or triggering custom actions through APIs (like giving refunds or updating accounts) might be limited. eesel AI is built to perform custom API actions and handle advanced ways of sorting and tagging based on all the details in a ticket.
- Cost as you grow: Lastly, think about the cost as you grow. While HubSpot has basic chatbot features on lower plans, getting to the more advanced AI stuff might mean upgrading to higher, more expensive plans. eesel AI uses a simple pay-per-interaction model. This can be a more cost-effective way to grow your AI support automation without paying extra fees for each agent, which some platform add-ons require.
Think of eesel AI as a solution that works alongside your current helpdesk systems (like Zendesk, Intercom, or Freshdesk, which are often used within a HubSpot setup). It offers a more powerful, flexible, and cost-effective way to get advanced AI support automation, working with your HubSpot setup instead of making you switch.
Ready to build a smarter Hubspot AI chatbot?
HubSpot’s Chatflows tool is a great way to start automating your customer conversations. With just a few steps setting up flows, adding actions, customizing the look, and connecting your **Knowledge Base,** you can build a chatbot that runs 24/7 and takes the pressure off your team.
But if you want deeper AI capabilities, more knowledge sources, better testing, and smarter automation, eesel AI offers a more advanced path. It connects with your current helpdesk, pulls from past tickets and internal docs, and helps your bot go beyond the basics.