Internal tools / Admin & orchestration
Account Provisioning Console
An internal approval console that turns a pending account request into fully provisioned, default-configured accounts — with one approval and a live view of every step.
Live demo — click around
8
2
147
3
11
Interactive prototype with representative sample data.
The challenge
Onboarding a batch of customer accounts into a tenant is repetitive, error-prone, and easy to get inconsistent: someone has to create each account, remember to apply the right default settings, and activate them — all while another person waits. Requests also need a gate (who's allowed to approve?) and a paper trail (what was approved, by whom, and did it actually succeed?). Doing this by hand doesn't scale and offers no visibility when a step silently fails.
Our approach
I built an approval console where account-provisioning requests arrive automatically (e.g. via webhook) and land in a Pending queue. An authorized approver reviews a request in a detail drawer and either rejects it with a reason or approves it — which kicks off a Step Functions orchestration that authenticates, creates the accounts, applies default settings, and activates them. The UI streams step-by-step progress live and records every action to a history and audit trail, so the same screen handles the decision, the automation, and the accountability.
How it works
Requests arrive and queue up
New account-provisioning requests are ingested (e.g. via webhook) and stored in DynamoDB, then surfaced in the Pending tab with tenant, account count, plan, requester, and time.
Only authorized approvers get in
Access is gated by Azure MSAL single sign-on on the frontend and an API Gateway custom authorizer on the backend, so every request to the ~14 Go Lambda handlers is checked before it runs.
Approver reviews in a detail drawer
Selecting a row opens the request detail, where the approver can reject with a reason or approve to proceed.
Approval triggers the orchestration
On approve, an AWS Step Functions state machine runs the four-stage flow: Authenticate, Create accounts, Apply defaults, Activate — auto-creating the accounts with their default settings.
Progress streams back live
The console shows the orchestration advancing through each of the four steps in real time, so the approver sees exactly where a run is and when a stage fails.
Everything is counted and logged
Status cards track Pending, Processing, Completed, Failed, and Rejected, and a history/audit trail preserves what happened for every request.
Tech stack
Results
The console replaces manual, one-by-one account setup with a single reviewed approval that fans out into an automated, observable orchestration — giving approvers a clear decision point, live progress on every provisioning run, and a full audit trail. Because provisioning applies default settings consistently through the same state machine every time, outcomes are more uniform and failures become visible instead of silent. (Portfolio case study; quantitative impact figures should be supplied by the project owner.)
Avg. approval-to-active time
Accounts provisioned per run
Manual setup time saved per request
Provisioning success rate
Requests processed to date
Metrics to be populated with the project owner’s real figures.
Ready to Start Your AI Journey?
Let's build AI solutions that save time, reduce costs, and help your business grow. Book a consultation today.
