IBM Cloud Pak for Integration only: Creating flows by describing your integration
If you are using App Connect Designer in IBM® Cloud Pak for Integration, you can create a flow by describing your integration scenario in English natural language. Embedded artificial intelligence (AI) services, which are available with IBM Automation Foundation assets (formerly Asset Repository), are used to facilitate this process. Your text entry is analyzed to recommend high-matching assets that you can use to create your own flow. This capability accelerates flow creation.
Recommended assets are available only with App Connect Designer 12.0.1.0-r1 or later. Existing event-driven or API flows, which are either preloaded IBM App Connect templates or manually saved assets in your Automation assets instance, are offered as recommendations.
Before you begin
You must have access to an Automation assets instance in which Designer AI features are enabled to provide asset recommendations. A user with the Automation administrator role typically creates the instance from the IBM Cloud Pak Platform UI and grants access to members of one or more teams. For information about creating an Automation assets instance, see Automation assets deployment in the IBM Cloud Pak for Integration documentation. For more information about enabling or disabling Designer AI features, see Enabling AI-powered features for use with Designer and Disabling the AI-powered features.
About the preloaded IBM App Connect templates
When Designer AI features are enabled, Automation assets automatically loads a set of IBM App Connect templates on initial access. These templates are added as a remote by synchronizing to a public GitHub repository that stores a set of IBM App Connect templates at https://github.com/ot4i/app-connect-templates/tree/cp4i-templates/resources.
The templates are listed (as assets) in the assets table on the Browse
Assets tab, with the Owner column set to IBM App Connect templates
. The Remotes tab
also displays an entry for these IBM App Connect templates in the
remote repository configurations
table.
By default, Automation assets automatically synchronizes with the remote GitHub repository at hourly intervals. If preferred, you can edit the IBM App Connect templates remote to change this synchronization setting. Only Designer API Implementation and Designer Event-driven Flow asset types are supported for this remote. For more information about configuring remotes, see Remotes.
app-connect-templates
) locally. For more information, see Remotes.How the AI services generate recommendations and flows
From your App Connect Designer instance, you can view the App Connect assets that are stored in Automation assets and can select these assets for use. Initially, only the preloaded IBM App Connect templates are included as assets, but you can use the embedded AI services to help you create customized flows and build up your catalog of assets.
To help identify templates that you might want to add as flows in your App Connect Designer instance, Automation assets is
enhanced with a Watson AI service or model, which is pre-trained with some of the most commonly used
utterances and template names. An utterance is a natural language query that describes a
specific intent or scenario; for example, Notify me about new Salesforce leads
.
When you type the text that is conveyed as an utterance, the utterance and all available templates
in Automation assets are sent as input to the AI service to find the best
match. The AI service uses Watson NLP and AI libraries to find a similarity between the utterance
and the names of all the templates, and can display up to three top recommendations in App Connect Designer for you to choose from.
As you begin to develop flows that are customized from the IBM App Connect templates, or flows that are created from scratch in your App Connect Designer instance, you can add them to Automation assets for sharing. Any flows that you add as assets are then included in Watson AI processing that occurs for recommended assets. For more information about sharing your flows as assets, see Adding a flow to the Automation assets catalog.
The following diagram illustrates how the AI service processes utterances to
generate recommended assets.
Creating flows
Learn how to generate recommended assets.