Updated on November 7, 2024
Salesforce is one of the world’s leading CRM and customer service organizations. Their Knowledge Base software, available with the Salesforce Service Cloud package, is also considered one of the best for keeping your customers updated.
However, it’s difficult for any customer to navigate your documentation and find a key answer. This is why customers are increasingly opting to chat with these documentation databases.
If you want to give your customers the same option, create a FAQ chatbot.
Thankfully, we have a method in which you can train a chatbot with your Knowledge Base without doing any coding.
In this tutorial blog, we’ll take you through setting up a Salesforce Knowledge Base chatbot through Kommunicate. If you haven’t signed up, you can register on this page.
Step 1: Go to the Kommunicate Dashboard
- You’ll be on the Kommunicate Dashboard page.
- Start by clicking on Bots (A).
Step 2: Start Creating a Bot
- You’ll be on the Bot Integration Page (A).
- Scroll down to find the available AI models.
Step 3: Choose between Open AI, Anthropic and Gemini
Open AI, Anthropic, and Gemini (from Google) are the leading AI companies worldwide. At Kommunicate, you can access the latest frontier models to automate customer service with increased efficiency.
- Start by clicking on Create a Bot and opening the Bot Integrations page. Scroll down to find the option to train your chatbot with:
- Open AI’s ChatGPT (A). You can create ChatGPT 4-o Mini bots, ChatGPT 4-o bots, and ChatGPT 3.5-Turbo Bots
- Google’s Gemini (B) – We currently support Gemini 1.5 Flash bots
- Anthropic’s Claude (C) – We support Claude 3 Haiku bots.
Step 4: Choose an AI Model for your Knowledge Base chatbot
- First, choose the specific model you want to use. For Open AI, we provide the following models.
- GPT 4-o-Mini
- GPT 4-o
- GPT 3.5-Turbo
- Then select the parameters you want to set for your model (A)
- Maximum Tokens – This is used to set the limit of tokens generated by ChatGPT for each request. (Or the number of words your chatbot will generate for each answer).
- Temperature – You can adjust this to increase or decrease the creativity of the answers from ChatGPT. A number closer to 1 will be more creative, and a number closer to 0 will be less.
We recommend that you keep the temperature at zero for customer service processes.
- You can also integrate your Open AI account with Kommunicate using an API Key.
- Click on Save and Proceed to save the changes and proceed.
Note: Your integration to Google’s Gemini and Anthropic’s Claude models will also work similarly. Just choose the right AI provider in Step 3.
Step 5: Choose a Knowledge Source for Your Bot
Now that you’ve chosen an AI model for your Knowledge Base. The next step is to choose the data source on which we’ll train it. In this case, we want to choose Salesforce Knowledge as our database.
- To start, go to the Bot Integration page and click Manage Bots on the left-hand side.
- Click on Go to Bot Builder to reach the flow designer for chatbots.
- Click on Knowledge Source on the left-hand side menu to select the data sources from which your bot should learn.
- On the Knowledge Source page, click on Knowledge Base (A).
Step 6: Select the Salesforce Knowledge Base
- On the Knowledge Base page, select Salesforce (A).
Step 7: Create a Salesforce Knowledge AI Chatbot
- This will take you to the Salesforce Integration page.
- Click on Connect to Salesforce. You’ll be taken to the Salesforce Login page.
- Enter your credentials
Step 8: Train Your Salesforce Knowledge Chatbot
- Your chatbot will automatically be trained on your Salesforce Knowledge base. Additionally, you must set up welcome messages with different intents and entities for better interactions.
- You can find the recommended methods to create the perfect chatbots with our detailed guide.
It’s easy to integrate Salesforce Knowledge Base with all the best AI models. Use our step-by-step guide to create a Knowledge Base chatbot with the best ChatGPT, Claude, or Gemini models.
At Kommunicate, we want to simplify the daily life of customer service and product managers so that they can focus on customer problems. That’s why our platform is fully no-code and requires 0 tech know-how from your end.
Want to build your own scalable AI customer support system? Talk to us today!
As a seasoned technologist, Adarsh brings over 14+ years of experience in software development, artificial intelligence, and machine learning to his role. His expertise in building scalable and robust tech solutions has been instrumental in the company’s growth and success.