Announcements
I've built a Power Automate scheduled cloud flow that generates a daily AI-powered email summary for a Tax partner at our firm. The flow runs at 5 PM ET and does the following:
/me/
This works perfectly for a single user — /me/ resolves to whoever owns the Graph connector connection, and the email goes to that person.
The challenge: I now need to scale this to 20-50 users across our Tax practice. Each user needs the flow to pull THEIR emails, calendar, and Teams data — not mine (the flow creator's).
What I've considered:
/users/{email}/
My questions:
Environment: Power Automate Premium, Copilot Studio, AI Builder (Claude Sonnet 4.6), Microsoft Graph connector, EisnerAmper tenant.
Any guidance or patterns from people who've solved this would be hugely appreciated!
In Power Automate, a shared scheduled flow cannot dynamically use /me per user — it always runs under the identity of the configured connection. This is by design and aligns with Microsoft’s consent and auditing model. For scalable automation across 20–50 users, the correct enterprise pattern is to use application permissions in Microsoft Graph with /users/{id} endpoints, looping through a managed user list (Dataverse or SharePoint). This requires admin consent but is the supported and secure approach. Client secrets should be stored in Azure Key Vault (recommended) or, where available, use Entra workload identity. Avoid storing secrets in SharePoint or environment variables for production scenarios. AI Builder costs can scale non-linearly based on user count, prompt volume, and token size, so guardrails like pre-filtering content and skipping low-activity users are essential.
/me
/users/{id}
In contrast, Microsoft Copilot Studio runs Graph calls using the end user’s delegated identity, which is why /me works there — but it does not support background scheduled automation. Recommended architecture: Use Power Automate (application permissions) for background daily summaries and Copilot Studio (delegated) for interactive queries. This hybrid model is the standard enterprise pattern for scalable, governed AI automation.
Under review
Thank you for your reply! To ensure a great experience for everyone, your content is awaiting approval by our Community Managers. Please check back later.
Jump in, show your community spirit, and win prizes!
Expanding mentorship, skilling, and AI innovation
These are the community rock stars!
Stay up to date on forum activity by subscribing.
Valantis 377
11manish 279
David_MA 234 Super User 2026 Season 1