Automating support: an (Open Source) AI chatbot for seamless participant engagement
Disclaimer: This article is not intended for commercial promotion. I accept no responsibility for any consequences arising from the use of the tools mentioned in this blog.
This blog is also translated into Dutch; click here. For other languages, you can select the flag below to translate it into your preferred language.
Recently, I was asked to deliver a workshop on AI trends, the future of AI, and how organizations can effectively use AI tools and programs. The workshop, designed for over 200 participants, would be followed by a self-learning phase, where participants could apply the knowledge gained. Additionally, I was asked to be available to answer any questions participants might have during this phase.
However, supporting over 200 participants while managing my other responsibilities would be extremely time-consuming and would significantly increase costs for the client. To resolve this problem, I started exploring ways to automate my expertise and provide ongoing support, not just for this workshop but for all future clients.
One solution I’ve been considering for some time is creating an AI chatbot. This chatbot would allow participants to get immediate answers to frequently asked questions, receive guidance on workshop topics, and seek assistance at any time—without requiring my constant involvement.
Here are my main requirements for the chatbot:
- Ease of use: The chatbot must be simple to install and maintain, without requiring advanced technical skills. It should be user-friendly so that it can be implemented by anyone with basic technical knowledge.
- Cost-effective: I believe solutions should be affordable. The chatbot should be budget-friendly to develop and maintain.
- Scalability: The chatbot must be capable of handling large datasets and managing multiple interactions simultaneously, considering the number of participants it will serve.
- Self-hosting: It’s important that the chatbot be installed on my own hosting server. This ensures full control over data security, customization, and functionality while avoiding reliance on third-party platforms.
- Custom knowledge base: The chatbot should be trained using my own knowledge, experiences, and frequently used reports. This way, it can answer common FAQs based on the materials I’ve developed, ensuring accurate and relevant responses.
After conducting thorough research, I wanted to share my findings with others who might be facing similar challenges. In this blog, I’ll provide an overview of available tools and platforms for creating an AI chatbot, focusing on those that meet these specific requirements. This blog isn’t a step-by-step installation guide, but rather a review of the possibilities.
AI Chatbot solutions overview
The table below highlights the options I found during my research. If you have any questions, don’t hesitate to reach out to me.
Keep in mind that connecting the AI chatbot to speech capabilities will involve additional costs beyond installation, hosting (if self-hosted), and any subscription fees, depending on the program you choose. You can add this speech option by using third-party Speech-to-Text (STT) and Text-to-Speech (TTS) services or by creating your own. The same applies for AI Avatar Chatbots, which I also highlighted in my previous posts. For more information, click here.
Open Source Frameworks/Platforms
| Framework/Platform | Voice Integration Options | Notes | Costs | Language | Installation Ease |
| Rasa | – Third-party STT services (e.g., Google, IBM) – Third-party TTS services (e.g., Google, Amazon) |
Requires custom implementation for STT/TTS | Free (self-hosted), costs for STT/TTS services | Python | Moderate |
| BotPress | – Web Speech API for web applications – Third-party STT/TTS services |
Custom modules can be created for integration | Free (self-hosted), costs for STT/TTS services | Node.js | Moderate |
| Haystack | – Custom STT/TTS integrations (varies) | Framework for building search-based chatbots | Free (self-hosted) | Python | Moderate |
| Rocket.Chat Chatbot (with Hubot) | – Custom integrations via Hubot | Open-source team collaboration platform | Free (self-hosted) | Node.js | Moderate |
| Chatwoot | – Web and messaging integrations | Customer support platform with chatbot capabilities | Free (self-hosted), paid plans available | Ruby | Easy (self-hosted) |
| Elasticsearch | – Custom STT/TTS integrations (varies) | Manage large datasets effectively with custom chatbot | Costs for hosting/maintenance | Java | Challenging |
| DialoGPT (OpenAI GPT-2/3) | – Custom integrations using HuggingFace Transformers | Pre-trained models for conversational AI | Free (self-hosted with HuggingFace), costs for API usage | Python | Challenging |
| Tock | – Third-party STT/TTS services | Multi-channel bot framework | Free (self-hosted) | Node.js | Moderate |
| DeepPavlov | – Custom STT/TTS integrations (varies) | Open-source conversational AI framework | Free (self-hosted) | Python | Moderate |
| ChatterBot | – Basic text input/output | Python library for generating automated responses | Free (self-hosted) | Python | Easy |
| Jasper | – Third-party TTS services | AI-powered writing assistant with conversational abilities | Free (self-hosted) | Python | Moderate |
| Clara | – Custom voice integrations | Open-source chatbot framework for various applications | Free (self-hosted) | JavaScript | Moderate |
| BotMan | – Third-party STT/TTS services | PHP framework for building chatbots | Free (self-hosted) | PHP | Moderate |
| Kuki (formerly Mitsuku) | – Custom integrations for messaging platforms | Award-winning conversational AI chatbot | Free (self-hosted) | JavaScript | Easy |
| OpenDialog | – Custom STT/TTS integrations | Conversational platform for creating complex dialogs | Free (self-hosted) | PHP | Moderate |
| Jitsi | – Voice and video integrations | Open-source video conferencing platform with chat capabilities | Free (self-hosted) | JavaScript | Moderate |
| Synapse | – Custom messaging integrations | Open-source matrix server for real-time communication | Free (self-hosted) | Python | Challenging |
Costing Frameworks/Platforms
| Framework/Platform | Voice Integration Options | Notes | Costs | Language | Installation Ease |
| Dialogflow | – Google Assistant integration | Built-in voice support; easy webhook connections | Pay-as-you-go pricing; free tier available | Google Cloud | Easy |
| Voiceflow | – Designed specifically for voice applications | Visual interface for Alexa/Google Assistant design | Subscription-based pricing; free tier available | Web-based | Easy |
| Amazon Lex | – Native voice input and TTS | Part of AWS ecosystem; easy integration with other AWS services | Pay-as-you-go pricing; free tier available | AWS SDKs | Easy (with AWS knowledge) |
| Flow XO | – Built-in integrations with various platforms | Automation platform with chatbot features | Subscription-based pricing | Web-based | Easy |
| Dialogflow CX | – Enhanced Google Assistant integration | Advanced version of Dialogflow with additional features | Pay-as-you-go pricing; free tier available | Google Cloud | Moderate |
| Joomla Chatbot Extensions | – Custom plugins for voice and chat | Extends Joomla CMS with chatbot functionality | Costs for plugins | PHP | Challenging |
| Microsoft Bot Framework | – Integration with Azure services | Comprehensive framework for building and connecting bots | Pay-as-you-go pricing | .NET | Moderate |
| IBM Watson Assistant | – Integration with various STT/TTS services | AI-powered virtual agent with various capabilities | Pay-as-you-go pricing; free tier available | IBM Cloud | Moderate |
| Tidio | – Messaging integrations | User-friendly platform for chatbot and live chat | Subscription-based pricing; free tier available | Web-based | Easy |
| ManyChat | – Integrations with Facebook Messenger and SMS | Marketing-focused chatbot platform | Free tier available, paid plans as needed | Web-based | Easy |
| Chatbot.com | – Integrations with various messaging apps | Platform for building chatbots across multiple channels | Subscription-based pricing; free tier available | Web-based | Easy |
| Zendesk Chat | – Integration with Zendesk support platform | Customer support chat platform with chatbot capabilities | Subscription-based pricing | Web-based | Easy |
| Landbot | – Web-based voice integrations | No-code platform for building conversational experiences | Subscription-based pricing | Web-based | Easy |
| Freshchat | – Messaging and voice integration | Customer engagement software with chatbot functionality | Subscription-based pricing | Web-based | Easy |
| PandaChat | – Voice and text messaging | AI-driven chatbot for customer support and engagement | Subscription-based pricing | Web-based | Easy |
| MobileMonkey | – Integrations with Facebook Messenger and SMS | Marketing chatbot platform with conversational capabilities | Subscription-based pricing | Web-based | Easy |
| Drift | – Messaging and voice integration | Conversational marketing and sales software | Subscription-based pricing | Web-based | Easy |
| Intercom | – Messaging and voice integration | Customer messaging platform with various integrations | Subscription-based pricing | Web-based | Easy |
| OpenAI GPT-4 | – Custom integrations using OpenAI API | Advanced conversational AI model for various applications | Pay-as-you-go pricing | Python | Moderate |
Why I chose Rasa: the ideal AI chatbot solution for customized learning
If you’re curious about my choice, I selected the Rasa Chatbot because it aligns perfectly with my requirements for creating a customized AI solution. Rasa offers a high degree of flexibility, allowing me to train the chatbot using my own knowledge, experiences, and frequently used reports. This means the chatbot can provide accurate and relevant answers to the most common FAQs, making it a valuable resource for workshop participants.
Moreover, Rasa enables easy self-hosting, giving me full control over data security and customization without relying on third-party services. Its user-friendly interface ensures that I can implement and maintain the chatbot without needing advanced technical skills. Additionally, Rasa is cost-effective, which is essential for ensuring that I can provide an affordable solution for my clients.
By using Rasa, I believe I can create a powerful tool that not only improves participant engagement during the workshop but also provides ongoing support as they navigate their self-learning journey. This makes it an ideal choice for my needs and those of my clients. However, time will tell. First, I need to move to the implementation phase, and after some time, I’ll be able to share more! 🙂
Demos of what can be created with Rasa can be seen at the following website:
To clarify the advantages and disadvantages of Rasa, here is a table with more information below.
| Advantages of Rasa | Disadvantages of Rasa |
| 1. Open Source: Free to use and customizable. | 1. Steep Learning Curve: Can be hard to learn at first. |
| 2. Custom Training: You can train it with your own data for better answers. | 2. Time-Consuming: Building a chatbot takes longer compared to easier tools. |
| 3. Self-Hosting: You can host it on your own server for better security. | 3. Needs Good Hardware: Requires decent computer resources to run well. |
| 4. Smart Understanding: Good at figuring out what users want. | 4. Harder to Connect: Linking it to other systems can be complicated. |
| 5. Community Help: Lots of users share tips and help online. | 5. Fewer Templates: Less ready-made designs, so you have to create more from scratch. |
| 6. Multi-Language Support: Works in different languages. | 6. No Visual Editor: No drag-and-drop tool for designing chats, which can be tricky. |
| 7. Keeps Context: Remembers what users talked about during the chat. | 7. Testing Can Be Tough: Figuring out if it works can be complicated. |
| 8. Scalable: Grows easily as more people use it. | 8. Basic Analytics: Limited tracking features; you need extra work for detailed stats. |
| 9. Custom Actions: You can add special features easily. | 9. Updates Can Be Tricky: Managing updates and software can be difficult. |
| 10. Regular Updates: New features and improvements come out often. | 10. Dependency Issues: Sometimes requires extra work to keep everything running smoothly. |
Tags: AI | Chatbot | Automation | KnowledgeManagement | SelfLearning | TechTrends | Workshop | FutureOfWork | DigitalTransformation | Innovation | AITrends | MachineLearning | EducationTechnology | Personalization | ClientEngagement | DigitalLearning | DataSecurity | OnlineTraining | TechnologySolutions | FutureSkills | InnovationStrategy | ChatbotDevelopment | SelfImprovement | KnowledgeSharing | SmartLearning | TechInnovation | UserExperience | LearningAndDevelopment | BusinessGrowth | AIChatbots | AutomationTools | OpenSource | Tech | Avatars | Server | CostSaving | CommercialUse | Integration | DataScience | Software | Programming | AIApplications | DigitalTools | BusinessTech
You may also like
Open-Source (AI) Software with reseller rights
Identifying websites with genuine Reseller Rights
