Secure Salesforce handoff to Copilot Studio with deployable samples
Diesen Beitrag auf Deutsch lesen
Updated integration guidance adds Apex samples, Named Credentials, permission setup and scripts for connecting Salesforce Einstein Bot with Copilot Studio.
TL;DR
Salesforce Einstein Bot forwards unmatched messages to Copilot Studio through Direct Line. Deployment scripts create three Apex classes, credential settings and permission grants, while Named Credentials keep Direct Line secrets out of code. The integration supports unauthenticated agents and uses configurable polling with a default five-second delay and five retries.
Original by Adi Leibowitz, on The Custom Engine. Read the original
This is our own summary, not a republication or full translation.
Governance takeaway
- Makers: Configure Einstein Bot Welcome and Confused dialogs so unmatched messages reach Copilot Studio and responses return to customers.
- Admins/CoE: Run the deployment script, then add the Direct Line secret manually to the External Credential because secrets are intentionally excluded from scripts.
- Security/Compliance: Use Named Credentials and Chatbot permission-set access to centralize secret rotation and restrict credential-principal access.
Frequently asked questions
How does the Salesforce Einstein Bot Confused dialog hand off messages to Copilot Studio?
The Confused dialog forwards an unmatched customer message through the Direct Line API, then displays Copilot Studio’s response in Einstein Bot.
What does the Salesforce Copilot Studio deployment script configure?
It deploys three Apex classes, creates callout and credential settings, grants the classes to the Chatbot permission set, and adds principal access.
Why should Salesforce Named Credentials store the Direct Line secret?
Named Credentials keep the secret out of Apex source, centralize rotation, and limit credential-principal access to the Chatbot permission set.