TL;DR If you are looking to work only with fine-tuning your intent detection algorithms go for Rasa. If you want to get an accurate natural language understanding chatbot on your website quickly and measure results, go for AlphaChat.
1. Introduction
Rasa.com and alphachat.ai are some of the most advanced Conversational AI tools for building Natural Language Understanding chatbots for customer service. In this post we will be comparing the two to make your choice between the two easier.
The focus is on understanding which tools suits best for those interested in customer service automation through natural language understanding smart chatbots.
2. Chatbot builder
rasa.com
- Slots as bot memory for storing data and referring to it later.
- Stories as examples of how conversations should go.
- Command Line Interface.
alphachat.ai
- No-code editor for building intents and workflows.
- Intents organized under folders to scale to 1,000+ intents.
- WYSIWYG chatbot text editor.
- Add links to other topics, links to URLs, images, carousels.
3. Training
rasa.com
- Open source NLP.
- See source code of the algorithms, adjust system parameters.
- Works out-of-the-box with BERT, HuggingFace Transformers, GPT, spaCy.
- Multiple intent handling.
- Measure F1 score and model confidence.
alphachat.ai
- Insert phrases under intents.
- Deep learning algorithms optimized for training speed and accuracy.
- Train with one-click.
- Training data organized under intents. Searchable and modifiable.
- Multiple intent handling.
- ReTrain function where AI suggests improvements into training data.
- Overlap function where AI shows overlap of phrases across intents.
4. Languages supported
rasa.com
alphachat.ai
- All European languages + Russian + Arabic.
5. Chat widget
rasa.com
- No widget available.
- Share your bot through Rasa X shareable URL for testing. Rasa X is an open source tool for analyzing user interactions.
alphachat.ai
- Fully configurable widget with customizable colors and logos.
- Shows text, images, videos, buttons, forms, carousels. Send files.
- Automatic customers language detection with URL parameters.
- Detects whether agents online or offline.
6. Metrics for measuring customer support efficiency
rasa.com
- No built-in customer support metrics.
alphachat.ai
- Measure self-service resolution rate.
- Measure direct costs saved.
- Customer satisfaction metrics.
7. Integration to RCS, WhatsApp etc.
rasa.com
alphachat.ai
- API available for integration.
8. Live chat availability
rasa.com
- No built-in live chat.
- Fallback action to assist human to 3rd party chat handover.
alphachat.ai
- Full live chat functionality built-in.
- Efficient handover from bot to agents with conversation history.
- Co-pilot mode where AI suggests answers to agents.
- Ability to connect to 3rd party customer support platforms (Zendesk, Salesforce).
9. Authentication of users
rasa.com
alphachat.ai
- Available through AlphaOS where custom code can be written into intents.
10. Enterprise features
rasa.com
- SSO
- Access Control
- Deployment Environments
alphachat.ai
- SSO
- Access Control
- Deployment Environments
- SLA. Dedicated support in Slack, Teams.
11. API
rasa.com
- HTTP API for training models, sending messages, testing.
alphachat.ai
- APIs for logs, statistics, history.
12. Price
rasa.com
- Rasa Open Source is for free.
- Rasa Enterprise pricing available on request.
alphachat.ai
- €399/month
- 1 NLU chatbot.
- 1 agent live chat license.
- Up to 3,000 chats / month included.
- Customizable widget.
- Deflection rate, customer happiness metrics
- Enterprise plan for API, AlphaOS, additional bots.
