5 different ways to create a chatbot using AI tools
With these tools, it’s easier than ever to create an intelligent AI-powered chatbot.
AI chatbots are like the Swiss Army knives of automation—they can handle customer support, streamline HR, handle questions within Slack, and even make your blog interactive.
But which tools should you use, and how do you get started? Let’s take a look at 5 fantastic chatbot tools to try—and some examples of what you can do with them.
1. Chatbase: Build a no-code chatbot to tackle anything you want
Chatbase is a versatile platform that makes chatbot creation ridiculously simple. Whether you need a bot for customer support, sales, or internal processes, this tool can handle it.
Chatbase is easy to use, has outstanding options for customization, and integrates cleverly through tools like Zapier and Slack.
Let’s quickly run through the steps to create a chatbot using Chatbase:
Sign in to your Chatbase account and click the “New Chatbot” button
Upload documents like product manuals, FAQs, or marketing materials. Chatbase will “read” them to build the bot’s knowledge base.
Check the chatbot has understood the info you uploaded, e.g. by asking a question from your FAQ document.
Customize the chatbot’s default instructions, tailoring them to your scenario (e.g. an AI chatbot to support employee onboarding). This often means giving the Chatbot specific constraints, like “No Data Divulge” and “Maintaining Focus”.
💡 Get the full course Build AI chatbots with Chatbase to see exactly how to create different types of chatbots for HR, lead generation, e-commerce, and customer support.
2. ChatGPT and Slack: Create an AI assistant directly into your Slack workspace
If your team lives on Slack, you could get a huge amount of value from an OpenAI-powered chatbot.
Zapier makes it easy to connect Slack with OpenAI (ChatGPT), creating a chatbot that can answer questions, provide summaries, or even brainstorm ideas.
Here’s how to create your AI assistant within Slack:
Set up a new Slack channel and give it a name like #chatgpt so that you’ve got a place to chat with the AI chatbot
Build a Zap that triggers whenever a message is posted in your #chatgpt channel. Make sure it doesn’t trigger bot messages, or the AI will end up stuck in an infinite loop, triggering itself every time it replies to itself—not ideal!
Next, add a Zapier step to extract the ‘Ts’ value from the trigger output (this allows for multiple ongoing chats, with AI responses and follow-up messages going in a Slack thread).
Add an action step to your Zap with the app “ChatGPT” and the event “Conversation”. Configure the message, bring in the user’s text, select the model to use (GPT-4o-mini is the fastest and cheapest), and set the “Memory Key” as the output “Ts” value from the previous step.
Finally, the Zap should send back the AI’s response as a Slack thread. You’ll need to use “Slack” as the app and “Append Entry and Schedule Digest” as the event.
Try out your new Slack assistant. You should be able to ask questions and get responses, just as you would with ChatGPT itself.
💡 For full step-by-step instructions, take a look at our tutorial Build your own ChatGPT in Slack.
3. Claude: Create a custom AI chatbot to streamline HR support
Let’s face it—nobody wants to sift through a giant HR handbook to check a detail about the PTO policy or to get one specific answer about the dental plan.
With Claude, you can build an HR chatbot that gives employees instant answers to their questions, using your HR documents as a knowledge base. Claude Projects is the perfect tool to create a custom chatbot that has a wealth of company-specific knowledge.
This is how you can create a Claude-powered HR chatbot:
Log in to your Claude Dashboard, click the Projects tab, and then click “Create Project”. Give your project a name and description: your team will see these, so make them clear and obvious.
Include project knowledge, by uploading all relevant HR documents. That might be your remote work policy, benefits FAQ, or your full employee handbook.
Use custom instructions to set how the chatbot should behave and what sort of answers it should give. For instance, you might say, “Your task is to accurately answer the question based solely on the information contained in the Remote Work Policy knowledge source.”
Make sure you instruct the chatbot on what to do if it can’t answer a question. For instance, you might get it to apologize and ask them to email HR.
Test the chatbot out, asking questions that are included in the information you’ve given it, to see if it answers these correctly. Try it out with questions that go beyond its information, too (e.g. “How do I set up a PIP?”) and check that it responds appropriately, by prompting you to contact HR.
💡 Get all the instructions for building your Claude HR chatbot in our detailed tutorial Create an HR chatbot for employees.
4. Zapier: Build a chatbot to make sure all copy meets brand guidelines
Zapier is a powerful tool for creating chatbots tailored to specific knowledge sources. One clever way you can use this within your team is to create a chatbot that can revise copy to match brand guidelines.
Here’s how to build your chatbot, using a knowledge source:
Login to Zapier and create a blank chatbot (or use one of the pre-build templates if you prefer).
Customize the chatbot’s greeting, setting the tone and explaining what it does. E.g. “I’m here to help with any questions about our brand guidelines.”
Connect Zapier and your OpenAI account to use GPT-4. You can use GPT-3.5 without an OpenAI account—but this won’t get such good results.
Head to the Instructions tab and upload your brand guidelines as a PDF (or any other knowledge source you want to use).
Then, set up actions for the chatbot, like giving a button so the user can copy its output. This lets users save and share what the chatbot comes up with, if it needs to rewrite their text to bring it in line with your brand guidelines.
Upload a sample piece of text (like a company announcement) and see how the chatbot provides feedback.
You can now embed the chatbot on your website, add it to a Zapier Interface, or create a secure URL where team members can access it.
💡 Get step-by-step instructions in our tutorial Build an AI chatbot with a knowledge source.
5. ElevenLabs: Make your blog posts interactive with a voice chatbot
ElevenLabs has a whole suite of tools plus an API that you can use to add an “agent” (AI voice chatbot) to your website or app.
Lots of companies use these for sales and support—but a clever way to use the agent is to add it to an in-depth article, letting users interact with your content and ask questions.
Here’s how to create an agent for your content:
Open the ElevenLabs dashboard and click “+” under the “Agents” tab. Name your agent something descriptive and choose the “Blank Template” option.
Tell the chatbot its purpose and explain that it should only respond using information from the article in its context window. You’ll need to upload the article by giving it the URL to it.
Test out the Agent, calling it and asking it questions related to the article. It should respond like a human would, listening, answering, and even handling interruptions.
Happy with your new AI agent? Go ahead and put it live on your website using the “Widget” tab options (customize the look and feel first, if you want).
💡 Find the full guide with all the instructions you need in Add a ‘talk to this article’ widget to your blog.
Why your business needs AI chatbots
AI chatbots can handle so many tasks. They’ll tackle questions from employees—and from potential customers—freeing up time. Using chatbot tools like Chatbase, ChatGPT, Slack, Claude, Zapier, and ElevenLabs, you can:
Create chatbots without the need for any coding
Answer questions automatically in Slack
Handle HR inquiries instantly and accurately
Make sure everyone sticks to brand guidelines
Add voice chat to your blog posts for stunning interactivity
Creating a chatbot doesn’t need to be difficult—so try out one of our tutorials today.