Table of content
At a glance
- What you can do: automatically receive a prioritized list of content opportunities (new content and optimizations) in a Teams channel, interact with the agent in the thread, and add topics to your editorial planning in one click
- Who it's for: content managers, SEO project leads, marketing directors
- Why it matters: your SEO insights arrive in your daily work environment on a schedule you define, no need to open Semji
What changes for you
Until now, Semji's AI agents existed exclusively inside the editor. Accessing your content opportunities required logging into the platform, navigating to the Intelligence Hub, and manually cross-referencing position data, search volumes, and competitor rankings — a time-consuming process that often got pushed back.
Before: you had to come find the information in Semji. SEO analysis remained a one-off exercise, disconnected from the team's day-to-day rhythm.
Now: the agent comes to you. At the frequency you choose (weekly, monthly, or custom), it automatically analyzes your Intelligence Hub data, identifies content to create and pages to optimize, and publishes a structured, actionable summary in your Teams channel.
| Before | With proactive agents in Teams |
|---|---|
| Manual login to Semji required | Insights arrive in Teams without any action on your part |
| Tedious cross-referencing of Intelligence Hub data | The agent automatically crosses positions, volumes, and competitor data |
| Semji perceived as a disconnected silo | SEO intelligence embedded where the team already collaborates |
| No recurring engagement loop | A regular SEO touchpoint anchored in the team's workflow |
Good to know ℹ️: this feature requires the "AI Agents" add-on and an active Intelligence Hub on your workspace, as well as a Microsoft Teams integration configured by an administrator.
What you'll get
At each scheduled run, the agent posts a structured message in the configured Teams channel with two sections.
- New content to create: a prioritized list of Content Gaps with, for each opportunity, a suggested title, primary keyword, search volume, top competing domain, estimated traffic gain, and difficulty rating
- Existing content to optimize: your pages sitting outside the top 10, with a semantic gap diagnosis and the recovery potential for each
- Action buttons: an "Add to Planning" button per opportunity to move a topic directly into Semji's editorial planning without leaving Teams
The agent then starts a discussion thread where it breaks down each recommendation in detail. You can reply to refine the analysis, shift to a different topic cluster, or ask follow-up questions.
How to use it
Setup happens in two phases: first on the Microsoft Teams side (by a Teams administrator), then on the Semji side (by a Semji administrator).
1) Install the Semji app in Microsoft Teams
From Microsoft Teams, open the app store and search for "Semji." The application is called "Semji for Microsoft Teams" and is also available on Microsoft Marketplace. Click "Get it now" then "Add."
Warning ⚠️: this installation can only be completed by a Microsoft Teams administrator. On first install, a Microsoft 365 tenant administrator must grant the requested Microsoft Graph permissions, including the ability to read channel messages without being mentioned.
2) Select the destination Teams channel
Once the app is added, select the Teams channel where the agent should publish its recommendations, then confirm with "Go." The app posts a welcome message in the channel with instructions for the next steps on the Semji side.
3) Connect the Teams integration in Semji
In Semji, navigate to Settings, then Integrations, under the Collaboration tab. Next to "Microsoft Teams," click "Connect." The connection wizard opens: click "Connect Microsoft Teams." A Microsoft authentication window (login.microsoftonline.com) appears to sign in with an admin account and grant OAuth consent.
Warning ⚠️: this step requires a Microsoft 365 administrator account on the client side. Make sure this account is arranged before starting setup.
4) Create and configure the agent in Agent Hub
Once the connection is established, open Agent Hub in Semji, go to the Agents tab, and find the "Workspace agents" section. Select the "Content Opportunities" agent. Configure the following settings:
- Destination Teams channel: select the channel connected in the previous step
- Execution frequency: weekly, monthly, or custom schedule
- Filtering criteria: strategic topic clusters, page types, content to exclude
Click "Save & Launch" to activate the agent. A manual test run is available to validate the format and relevance of recommendations before the first scheduled execution.
Best practice ✅: always run a manual test before enabling the schedule. This lets you verify that recommendations match your target topics and that the message format displays correctly in Teams.
5) Interact with the agent from the Teams thread
At each run, the agent opens a discussion thread in the configured channel. You can reply directly in that thread to refine the analysis ("Filter for the 'Young driver' persona"), ask follow-up questions ("Why does this competitor dominate this keyword?"), switch to a different topic cluster, or drill deeper into a specific opportunity. The agent retains full conversation context throughout the thread.
Tips for best results
Match the frequency to your editorial rhythm. A monthly run suits a strategic review in a leadership meeting. A weekly frequency works better for an editorial team in active production that wants to factor SEO opportunities into its content sprints.
Create a dedicated Teams channel rather than posting to your team's general channel. This makes it easy to browse the history of recommendations and prevents the agent's messages from getting lost in day-to-day conversation.
Set topic filters from the very first configuration. An agent without filters will push every Intelligence Hub opportunity, which can create noise. By targeting your workspace's strategic topic clusters, every recommendation delivered in Teams is immediately actionable.
Pro tip 💡: after a few weeks of use, look at how many opportunities get added to planning directly from Teams. That's the clearest signal for tuning your filters and execution frequency.
Go further
Once the agent is live, you can combine this feature with other platform capabilities to maximize your editorial output. From Semji's editorial planning, find every topic added via Teams and assign them to your writers. From the content editor, activate AI+ Content or Atomic Content to generate drafts directly from the brief the agent surfaces.
The agent also supports free-form instructions in the Teams thread. You can ask it to go deeper on a specific opportunity, surface the top three questions being searched around a keyword, or redo the analysis on a different topic cluster.
Warning ⚠️: each scheduled execution consumes 10 AI credits. Each agent response in the thread also consumes 10 credits. The feature requires the "AI Agents" add-on and an active Intelligence Hub on the workspace. Without position and competitor data in the Intelligence Hub, the agent has no data source to analyze.
Useful links:
- [Link to "Intelligence Hub: setting up your topic clusters and competitors"]
- [Link to "Editorial planning: plan and manage your content production"]
- [Link to "Agent Hub: discover and configure Semji agents"]