DataBot enables users to ask questions about data and visualize it using natural language.
ARCHITECTURE
DataBot integrates your data with your website/app without the need to copy your data to our servers. Your data remains securely stored in your own data warehouse.
DataBot is self-hosted, ensuring complete isolation. Moreover, as no data is exposed to LLMs, it provides one of the most secure conversational analytics solutions available.
If you prefer not to host the app yourself, you can use the SaaS version. Your data will remain securely stored in your own data warehouse.
Supported Databases
Supporting all major SQL and NoSQL databases
Athena
Oracle
SQL Server
PostgreSQL
MySQL
MongoDB
Snowflake
BigQuery
ClickHouse
Redshift
Data Governance & Security Measures
DataBot is designed with data security as a top priority, employing a comprehensive set of measures to protect your data, as outlined below.
Authentication
Username/password Authentication and Email Verification act as the first layer of security.
Email White-listing
If enabled, new users need to be granted access by an admin to be able to use the app. This adds a second layer of security.
No data Exposure to LLMs
Only the database schema (table and column names) is provided to LLMs, NOT the data itself.
Data Obfuscation
Using Dynamic Data Masking, admins can obfuscate some columns for certain user groups.
Fine-grained Access Control
Admins can manage the table-level and row-level access of users.
Data ExplorationUsers can ask questions about data using natural language.
VisualizationUsers can ask DataBot to visualize the results (e.g. Bar chart, Line Chart, ...).
Personal DashboardEach user has a personal dashboard that can be maintained using natural language.
Context AwarenessDataBot answers the user's question taking into account their previous questions in the chat session.
Data ObfuscationAdmins can mark some columns to be obfuscated for some user groups. Obfuscation is done using dynamic data masking.
Embedding in Your Website or App
Fine-grained Access ControlAdmins can manage users' access at database-level, table-level and row-level.
Two-Column FormatIn this format, the chatbot is displayed on the right, while a custom dashboard is shown on the left. The dashboard can be used to display database tables, giving users an idea of the data available for querying. It can also be used to present KPIs and charts. Row-level access controls are applied to the dashboard as well.
Data VerificationUsers can ask DataBot to explain how it generated the result to verify data accuracy.
Voice Transcription
Public SharingUsers can share the result publicly using a link. Admins can enable/disable this feature.
White-LabellingYou can replace the DataBot and Intellimenta logos in the app with your own logo.
Salesforce IntegrationIntelliSales is our Salesforce app that has a similar functionality to DataBot but also enables users to automate certain Salesforce operations using natural language. In the Enterprise plan you can use both Databot and IntelliSales, seamlessly integrated together.
Almost any question as long as it can be answered using the connected data. Note that the engine answers with data. So, instead of asking "what caused the sales increase?", it would be better to ask "show me top 3 best performing products in terms of total sales units". You can also issue commands for visualizing the data or managing your personal dashboard.
Regarding data exploration commands, the users question is translated into a language that the database understands. Using the translated command the data is fetched and shown to the user. Regarding visualization commands and personal dashboard, we have integrated ChatGPT with a BI tool to handle these commands.
Yes. Data is directly fetched from your database in real-time.
You can use the "explain" command to ask the engine explain its answers. This can help you verify the results.
Please see the "DATA GOVERNANCE & SECURITY MEASURES" section above. The self-hosted version provides the highest possible level of security.