Cognigy.AI 2.7 released

3 min read
Media contact: Cognigy February 25, 2018

COGNIGY.AI 2.7 has been released with more amazing features to help you create amazing conversational AIs. See below for our top 4…


1. Support for ALL languages
You read right! We now support ALL languages through the introduction of our new Generic Language option.
This option allows you create Flows for any input language, no matter if it’s Arabic, Russian, Japanese or even Klingon (no kidding!).
Flows built in the Generic Language still have full access to all Flow features and slots and become even more powerful through the next feature – Rule Intent.
Head over to the documentation site for more information.

2. Rule-based Intent Mapping
COGNIGY.AI’s machine learning-based intent mapper is one of the strongest in the industry, but what if you want to simply say “all utterances that include the word ‘order’ and a slot of type ‘food’ should be identified as intent ‘orderFood’“?
This is now possible with our new Rule Intents that sit alongside the regular ML Intents. You can create any number of CognigyScript based rules that will be executed against the user input and if the rules evaluate to TRUE, the intent is found.
Check the documentation for more.

3. Public Analytics Endpoint
As of this version, you can access your Flow analytics from our public analytics OData endpoint at
Using OData, you can easily integrate your Flow analytics data into Excel, PowerBI or other analytics tools of your choice. Common OData features, such as $filter, $select, $top, $skip, etc, are fully supported.
Please check the documentation for details on authentication, etc.

4. Improved SAY Nodes
In previous versions, the output of a SAY Node was randomly selected from all provided sentences. You now have three options:
– Random
Picks a random sentence from the provided ones
– Linear Execution
Executes the sentences in the order they were provided, one by one, each time the Node gets hit. It stops at the last element and keeps repeating it.
– Linear Loop Execution
Works like the Linear Execution, but starts at the first element again once the end of the list has been hit.
Albeit a small change, it makes conversation much more natural (“hi”, “oh, hi”, “hi”, “well, we already said hi”, “hi”, “ok, not sure why you keep repeating that, but hello again“).