Before
Reviewers had to update the Shadow IT registry for each decision and manually confirm whether the change was accepted. Invalid requests and failed writes required additional follow-up to determine what happened.
How it works
- Receive the decision request and read the application index.
- Validate the token, decision type, and application, then reject requests that fail any check.
- Store valid decisions and return a distinct success or write-failure response.
Workflow details
- Request intake: A secure webhook accepts a decision submitted in the request body or query parameters, including the application, proposed decision, submitter, and validation token.
- Validation: The workflow normalizes the request, checks the token, confirms that the decision is allow, tolerate, or block, and verifies that the application exists in the Shadow IT index.
- Decision recording: Valid requests are formatted as a Shadow IT decision record containing the application, decision, submitter, timestamp, and previous AI decision, then sent for storage in Artifacts.
- Outcome handling: The caller receives a success response when the record is saved, a refusal message when validation fails, or a separate write-failure response when storage does not accept the decision. The website demonstration replays synthetic read and write responses locally.
What changed
Every submission now passes through the same automated validation and storage path, with separate responses for invalid requests and rejected writes; reviewers receive the outcome directly instead of checking the registry and storage result manually.
Work removed
Manual registry entry, repeated application lookups, checking decision validity, and manually following up to confirm whether a write succeeded or failed.
Expected outcome
Each valid submission receives one explicit save outcome, removing at least two recurring manual actions per decision: registry updating and confirmation follow-up.
