Build your own Virtual Assistant!
The SDK to build customized
(voice-operated) Virtual Assistants
A virtual assistant can respond to questions or commands formulated in our own natural language. At the core of a virtual assistant is the software that interprets the question or command to an instruction a computer can understand.
With our SDK you can build a text-based Virtual Assistant that is tailored to your own knowledge-base. Add a commercial voice recognition software and you have a customized voice-operated Virtual Assistant!
Get a better Virtual Assistant!
Build on your knowledge base
Retrieve answers from your SQL databases
Your Virtual Assistant will translate natural language to SQL - enabling you to retrieve information from your own databases.
Intelligent
Understands the question - no guessing!
The DialogueTech Virtual Assistant will understand the meaning of a question - resolving ambiguities, understanding different formulations for the same question etc.
Precise
Finds a correct answer
Through the built-in linguistic capabilities and the added customization, you can ensure that each question generate a unique and correct answer or command.
Multi-lingual
Port to new languages
The SDK is built with Plug-in grammars and allow a high degree of reuse of applications when porting to new languages.
Easy-to-use SDK
Easy to build and maintain
Powerful tools for application building and maintenance keeps development and maintenance cost at a minimum
Local or hosted
Works on servers and mobile devices
Your application can be compiled to a number of different environments. In most cases it will be hosted but it can also run locally on for example Android devices - practical when internet coverage is unreliable.
Realise value from your Virtual Assistant!
Rationalize your customer service
With our technology you can build self-service that really works - providing intuitive access to all frequent questions. In most applications you can expect first-time resolution in more than 80% of all customer queries.
Build voice-control into products
Build functioning voice-control of e.g. cars (enabling eyes on the road, hands on wheel) or any other device where the user benefits from using his/her voice without being restricted to predefined commands.
Build natural language access into S/W products
Many software products contain vast amount of data in databases (customers, products, sales, etc., etc.) Make the data intuitively accessible – let the users ask for their desired information in free format!
Innovate new applications!
There are certainly many more applications we have not yet thought of – but within your market and with your current offering many of you will find new applications (Q&A-based manuals, automated premium-calling services, etc.)
Your own Virtual Assistant in 4 steps!
1. Define the domain
The "domain" is the information that you want to make accessible for your users.
If your knowledge is stored in a database you can skip this stage - the database is your domain!
If you e.g. have an FAQ you store answers in a database and define standard questions to address the database
Low cost: reuse of current repositories
2. Create domain model
A domain model is a representation of your information. In our case it is build on entities and relationships.
The steps to build a domain model are: Define entities (name of columns in database), define their relationships, add SQL to entities and add specific lingustic information
Precise: Will be able to retrive accurate answers by creating a unique SQL statement
Multilingual: Domain models can be transferred between languages (with minor modifications)
Low cost: Easy-to-use web-tool
3. Compile the domain model
Use our hosted compilation routine - just bring your file to our compilation routine and press "enter"! We will add linguistic libraries (grammar, dictionaries, semantic rules etc.) to you domain model and then build an executable file for you.
Intelligent: General language understanding is added and tools to analyse the questions/commands
Low cost: A few clicks and it is done!
4. Integrate and test
Integrate the executable file you receive from us with your own software - for example voice recognition software. You can see examples from our sample applications.
When you have used our application for a while you can use our Analyzer tool and log files to see where the application need improvement. Add entities, relationships etc. to enhance coverage and precision
Intelligent: During testing and maintenance the ability to handle customer requests will be fine-tuned
Low cost: Powerful tools for analysis, low cost to extend domain model