Power Platform Community Forum Thread Details

Hello Community,

I am exploring the feasibility of building an internal AI-assisted solution using Microsoft 365 tools for managing my company’s LinkedIn content, and I would appreciate expert guidance on what is realistically achievable today.

Scenario:
I want to create an internal assistant that:

- Generates daily LinkedIn post drafts based only on my company’s content (website, previous posts, brand guidelines, festivals, events, etc.)
- Can also generate post drafts based on specific topics provided by a user or team (for example, Microsoft 365 updates, company initiatives, or predefined content themes)
- Sends the generated post to a user or team for approval (approve / reject / edit)
- Publishes the post to the company’s LinkedIn Page only after approval
- Can run on a daily schedule (e.g., every morning at 9 AM)
- Optionally uses analytics from the company’s own LinkedIn Page (engagement, likes, comments) to improve future post suggestions
- Can generate event- or festival-based posts when an event occurs

Constraints:
- Prefer Microsoft 365 tools.
- Human approval before posting is mandatory

My questions:
1. Is this scenario fully or partially achievable with current Microsoft 365 and LinkedIn capabilities?
2. Which parts are realistically possible today, and which are not?
3. Can Copilot Studio be used effectively as the “assistant” in this workflow, with Power Automate handling approvals and posting?
4. Would using LinkedIn’s official APIs or third-party social media tools be required for posting and analytics?
5. Which components typically require paid licenses versus what can be done with standard M365 capabilities?

Thanks in advance for your insights.

Categories:

Autonomous agents

Hi there,

Great use-case — and the short answer is yes, this is achievable today with Microsoft 365 + Copilot Studio + Power Automate, with a couple of caveats around LinkedIn API surface. Let me address your questions one by one:

1) Is this scenario fully or partially achievable?

It's largely achievable. The official Microsoft LinkedIn V2 connector supports posting to LinkedIn and is available in Power Automate, Copilot Studio, and Logic Apps as a Standard connector (LinkedIn V2 connector reference).

2) What's realistically possible today vs. not?

Capability	Today	Notes
Daily scheduled draft generation	✅	Power Automate scheduled (recurrence) trigger + AI prompt
Grounded on company content (website, brand guidelines, previous posts)	✅	Use Copilot Studio agent with knowledge sources (file upload, SharePoint, web pages)
Topic-based drafts (user/team supplied topics)	✅	User input via Question node or Teams form
Human approval (approve/reject/edit)	✅	Power Automate Approvals (Create and test an approval workflow) — natively integrates with Teams/Outlook
Publish to company LinkedIn Page after approval	✅	LinkedIn V2 connector — "Share a company update V3" / similar actions
Scheduled daily run	✅	Recurrence trigger in Power Automate
Event/festival-based posts	✅	Use a Dataverse/SharePoint events table + scheduled flow
LinkedIn engagement analytics back into the agent	⚠️ Partial	LinkedIn V2 connector exposes limited analytics actions; deeper engagement metrics typically need direct LinkedIn Marketing Developer Platform API access (separate developer approval)

3) Can Copilot Studio be the "assistant" with Power Automate handling approvals and posting?

Yes — this is the ideal pattern:

Copilot Studio agent = the conversational front-end + generative draft author. Use agent flows / autonomous agents for scheduled/triggered execution (Agent flows and workflows overview – Microsoft Learn).
Power Automate = the orchestration plumbing (schedule, approval, LinkedIn post call, retry/logging).
LinkedIn V2 connector = the publish step, called from Power Automate after approval is granted.

4) Do you need LinkedIn's official APIs or third-party tools?

For posting to a Company Page, the LinkedIn V2 connector is sufficient and uses LinkedIn's official APIs under the hood. For richer analytics (impressions, follower demographics), you'd either need:

A LinkedIn Marketing Developer Platform–approved app + custom HTTP calls, OR
A third-party tool (Hootsuite, Sprout Social, etc.) that already has elevated API access.

5) Licensing

Standard M365 can cover Outlook approval emails, Teams notifications, and basic flows.
Power Automate Premium is required for premium connectors and for agent flows in some scenarios — check the Power Automate licensing guide.
Copilot Studio requires its own licensing for agents and message capacity (Security and governance – Microsoft Copilot Studio).
The LinkedIn V2 connector is Standard (no premium license needed for the connector itself).

Suggested architecture for your POC:

SharePoint document library → brand guidelines, past posts, content themes (knowledge source).
Dataverse table (or SharePoint list) → events/festivals calendar.
Copilot Studio agent with knowledge sources + scheduled trigger.
Power Automate flow: generate draft (AI Builder / Copilot prompt) → Approval action → on approve, call LinkedIn V2 "Share a company update" → log result.
Optional: weekly flow that pulls LinkedIn engagement and writes back to a Dataverse table the agent uses as additional knowledge.

Hope this gives you a solid roadmap!

Raghav Mishra — LinkedIn | PowerAI Labs

Found this helpful? Please mark ✅ "Does this answer your question?" so others searching for the same issue can find it quickly. A 👍 on "Was this reply helpful?" or a ♥ Like is also much appreciated!