Migrating from Cognigy.AI v3 to v4

The launch of Cognigy.AI v4 brought a plethora of new features, capabilities and UI improvements for conversation designers as well as developers. Under the hood, v4 advanced with revamped internal structures and models to keep Cognigy.AI the leading Conversational AI platform. 

These changes require Projects on v3.6 environments to be migrated to v4 and become Virtual Agents. Follow the procedure below for a self-governed step-by-step migration or contact Cognigy Professional Services to assist with the process.

1) Export Project from Cognigy.AI v3

Log in to Cognigy.AI v3 and use the Export function in the top-right toolbar (three dots menu) to create a ZIP file that bundles all Project resources. Store the export.

mceclip0.png

For later steps, you should note down the name of the respective Project / Virtual Agent and the relevant Locales (NLU languages).

2) Get an API Key for Cognigy.AI v4

Go to Cognigy.AI v4 to retrieve an API Key. If you don't have your own Cognigy.AI instance, go to signup.cognigy.ai and create a trial account. Access "My Profile" on the top-right and create a Key by clicking the plus sign in the API Keys section.

mceclip1.png

3) Use the "Cognigy v3 to v4 Migration-Helper" 

The Cognigy v3 to v4 Migration-Helper is a CLI tool that processes v3 exports and imports them directly into a Cognigy.AI v4 environment via API. The CLI is available for Windows, Linux and Mac. 

Go to Github for the full documentation and to download the tool:

https://github.com/Cognigy/Migration-Helper

All you need is the ZIP export, the API Key and the host URL to the target API server. Other options are available for naming and finetuning the process. Here is an example command: 

./migrator-linux -a YOUR_API_KEY -h https://api-trial.cognigy.ai -f EXPORT.ZIP

4) Review the error log file

The migration helper creates a log file that protocolls any errors during automated migration. You should store and review the error log file to identify any potential glitches in the process.

5) Finetune your Virtual Agent configuration

Before sending your Virtual Agent into production, some resources need to be updated manually as concepts have changed from v3 to v4.

These resources can be migrated:

  • Flows
  • Flow Nodes (except for Custom Modules)
  • Flow Intents
  • Lexicons
  • Playbooks
  • NLU Connectors
  • Processes (will be converted to Flows)
  • Endpoints

The following can't be migrated:

  • Custom Modules (are now Extensions and must be re-built)
  • Connections (were a different type of connection in v3)
  • Secrets (are now Connections in v4)

Done!

Your v3 Project is now available in Cognigy.AI v4 as a Virtual Agent.


Comments

0 comments

Please sign in to leave a comment.

Was this article helpful?
0 out of 1 found this helpful