AI Content Idea Agent with Approvals

Build an AI agent that finds content ideas and routes them through human approval.

Introduction

30 Second Summary

Every content creator knows the struggle of staring at a blank page, trying to come up with fresh ideas. You end up spending hours browsing articles and social media, only to forget half the good ideas you found.

In this project, you will build an AI-powered content idea pipeline in Microsoft Copilot Studio that discovers ideas from the live web, saves them to a raw file, and lets you approve ideas with a single button click in Excel that copies them into a structured masterfile.

What You'll Build

You type a topic into the test chat, your AI agent searches the web for content ideas, and you watch approved ideas land in your masterfile complete with Content IDs and format tags.

By the end of this project, you'll have:

An AI agent in Copilot Studio that searches the live web for content ideas on any topic you ask about.
An automated agent flow that captures discovered ideas into an Excel raw file with topic, format, and source URL.
A one-click button flow that lets you approve ideas directly in Excel, auto-generates a Content ID, and copies them into a structured masterfile with duplicate protection.
Secret Mission: Add AI pre-scoring that ranks ideas by relevance before you review them.

Are there any prerequisites?

You need a Microsoft 365 work or school account with access to OneDrive. A free Copilot Studio trial account is all you need to build and test your agent. Step 1 walks you through the trial signup.

Before We Start

Before we dive in, let's lock in what you're about to build and why it matters. You're creating an AI-powered content idea pipeline that automates the first stage of a content production workflow: discovering ideas from the web, filtering them through human judgment, and organizing approved ideas in a structured masterfile.

Set Up Copilot Studio and Excel Files

Your content pipeline needs two foundations before you can start building. First, a workspace where your AI agent lives and can be tested. Second, structured data files to catch the ideas it discovers.

In this step, you'll sign up for the Copilot Studio trial and create two Excel workbooks on OneDrive. These files become the raw ideas inbox and the approved masterfile that your agent writes to automatically.

In this step, get ready to:

Sign up for the Copilot Studio trial.
Create two Excel workbooks on OneDrive with formatted tables.
Verify both environments are ready.

Sign up for the Copilot Studio trial

Copilot Studio is the low-code platform where you'll build and test your AI agent. The trial gives you full access to create agents and test them in the built-in chat panel.

Open your browser and navigate to the Copilot Studio trial signup page: https://go.microsoft.com/fwlink/?LinkId=2107702
Enter your work or school email address (the same account tied to your Microsoft 365 license).
Select Next.
Follow the on-screen instructions to complete the signup process.

After signup completes, you land on the Copilot Studio Home page at copilotstudio.microsoft.com. You should see a description box prompting you to describe what your agent should do.

What can I do with the trial?

The trial lets you create agents and test them using the test chat panel. You cannot publish agents on the trial, but testing is all you need for this project.

Testing an agent from the test chat does not consume Copilot Credits. The trial environment expires after 30 days but can be extended.

🙋‍♀️ Can't complete the signup?

Make sure you're using a work or school email address, not a personal Microsoft account (like @outlook.com or @hotmail.com). The trial requires a Microsoft 365 work/school license.

If you see a message that you already have access, your organization may already have Copilot Studio provisioned. Try signing in directly at copilotstudio.microsoft.com.

Help me troubleshoot Copilot Studio trial signup issues.

Create your Excel files on OneDrive

Your pipeline needs two Excel files: one for raw ideas (everything the agent finds) and one for approved ideas that pass human review. Both must have their data formatted as proper Tables for the connector to recognize them later.

Open onedrive.com in your browser.
Sign in with your work account (the same one you used for the Copilot Studio trial).
Select + New near the top of the page.
Select Excel workbook.
A blank workbook opens in Excel for the web.
Type Idea in cell A1.
Type Source URL in cell B1.
Type Topic in cell C1.

Type Format in cell D1.
Type Date Found in cell E1.
Type Status in cell F1.
Type Content ID in cell G1.
Type Date Approved in cell H1.
Select all cells from A1 to H1.
Select Insert in the ribbon.
Select Table.
Confirm that My table has headers is checked.
Select OK.
Select the filename at the top-left of the page (it shows something like Book).
Type Content-Ideas-Raw and press Enter.

You should see your headers styled as a table with filter arrows on each column and alternating row banding. The Status, Content ID, and Date Approved columns stay empty for now. Your approval flow will fill them later. Excel for the web auto-saves, so your file is already stored on OneDrive.

Now create the masterfile using the same process.

Navigate back to your OneDrive files list (select the OneDrive logo in the top-left corner or use the browser back button).
Select + New.
Select Excel workbook.
Type Content ID in cell A1.
Type Idea in cell B1.
Type Format in cell C1.
Type Source URL in cell D1.
Type Topic in cell E1.

Type Status in cell F1.
Type Date Approved in cell G1.
Type Publish Date in cell H1.
Select all cells from A1 to H1.
Select Insert in the ribbon.
Select Table.
Confirm that My table has headers is checked.
Select OK.
Select the filename at the top-left and rename it to Content-Masterfile. Press Enter.

Your masterfile now has a formatted table with eight columns. The Publish Date column is for you to fill in manually after you schedule or publish the content. Only ideas that pass your approval will end up here.

Why does the data need to be a Table?

The Excel Online connector in Power Automate only recognizes data formatted as a Table (created via Insert > Table). If you just type headers into cells without applying Table formatting, the connector won't detect your data when you try to add rows later.

You can confirm a range is a Table by looking for filter arrows on the headers and alternating row colors.

Verify your setup

Before you move on, take a moment to predict what you expect to see. Do you think both the Copilot Studio Home page and your two Excel files are ready?

Open a new browser tab and go to copilotstudio.microsoft.com.
Confirm you see the Home page with the description box for creating an agent.
Switch to your OneDrive tab and navigate back to your files list.
Confirm both Content-Ideas-Raw.xlsx and Content-Masterfile.xlsx appear in your file list.
Open each file and verify you see the table formatting with filter arrows on every header column.

You should see the Copilot Studio Home page with its agent description box, and both Excel files in OneDrive with properly formatted tables. This confirms your environment is ready.

Don't see the description box on the Copilot Studio Home page?

If you don't see the box prompting you to describe what your agent needs to do, your environment may not support agent creation using natural language. Try selecting Create an agent under Start building from scratch instead.

If your Excel files don't show filter arrows on the headers, you may have skipped the Table formatting step. Select the header row, go to Insert > Table, and confirm the headers checkbox.

Help me verify my Copilot Studio and OneDrive setup.

Your workspace is ready. Next up, you'll create your first AI agent in Copilot Studio and teach it to search the web for content ideas.

Create Your Content Idea Agent

Your Copilot Studio environment is ready and your Excel files are waiting on OneDrive. Now it's time to build the brain of your content pipeline: an AI agent that searches the live web for ideas.

Right now, finding content ideas means manually browsing blogs, social media, and news sites. Your agent will do that research for you, pulling trending ideas from across the web in seconds.

In this step, get ready to:

Create an AI agent in Copilot Studio with a content-research description.
Enable web search so the agent can pull live content from the internet.
Test the agent to confirm it returns relevant ideas with source links.

Create the agent

Copilot Studio lets you create an agent by describing what you want it to do in plain English. The platform auto-generates the agent's name, instructions, and configuration from your description.

Return to the Copilot Studio Home page at copilotstudio.microsoft.com.
In the description box on the Home page, type the following description (up to 1,024 characters are allowed):

Help content creators find trending content ideas by searching the web. When asked about a topic, search for recent blog posts, articles, and social media trends, then summarize the top ideas with source URLs.

Press Enter or select the submit button to create the agent.
Wait for the agent to provision. You land on the Overview page once it's ready.

Copilot Studio auto-generates the agent name, description, and instructions from the text you provided.

Select the agent name field at the top of the Overview page.
Replace the auto-generated name with Content Idea Scout.
Select Save at the top of the page.

You should now see Content Idea Scout displayed as your agent's name on the Overview page.

What just happened?

Copilot Studio took your plain-English description and created a fully configured agent with generated instructions. The agent knows its purpose (finding content ideas) and how to behave (searching the web, summarizing results, including sources).

You renamed it to Content Idea Scout so it's easy to identify throughout this project.

🙋‍♀️ Don't see the Overview page?

If the agent creation seems stuck or you see an error, refresh the page and try submitting the description again. Make sure your description is under 1,024 characters.

help me troubleshoot agent creation in Copilot Studio

Enable web search

Your agent needs access to live web content to find current ideas. The Web Search setting lets the agent query Bing for public websites and return summarized results with citations.

On your agent's Overview page, look for the Knowledge section.
Confirm that the Web Search toggle is enabled (it should be on by default for new agents).

If Web Search is already toggled on, you're all set. If it's off, follow the steps below to enable it.

Select Settings in the top navigation.
Select Generative AI in the left menu.
Turn on the Use information from the web toggle.
Select Save and return to your agent's Overview page.

Confirm the Web Search toggle now shows as enabled in the Knowledge section of the Overview page.

How does web search work?

When Web Search is on, the agent generates a focused keyword query from the user's question and sends it to the Bing search service. Bing returns relevant results (titles, snippets, and links), and the agent synthesizes a unified answer that cites the sources.

This means your agent pulls live, up-to-date information from across the public web every time you ask it a question.

Test the agent

Time to see your agent in action. The Test your agent panel lets you chat with the agent in real time, exactly as an end user would.

Select Test at the top of the page to open the Test your agent panel.

Before you type your first message, what do you think the agent will return? Will it give you generic ideas or pull specific articles from the web with real links?

In the text field at the bottom of the panel, type the following and press Enter:

Find me 3 content ideas about sustainable fashion

The agent responds with ideas pulled from current web content, including source references and links. Each idea should be a distinct, actionable content suggestion grounded in a real article or post it found online.

Try another topic to confirm the agent works across different subjects. Type the following and press Enter:

What are trending content ideas about AI tools for small businesses?

Confirm the agent provides relevant, web-sourced responses for this topic too. You should see different ideas with different source links each time.

What should I see?

The agent should return a numbered or bulleted list of content ideas, each with a brief description and a source URL or citation. The ideas should be specific and current, not generic suggestions.

If you see citations or "Learn more" links in the response, that confirms web search is working and the agent is grounding its answers in real web content.

🙋‍♀️ Agent not returning web results?

If the agent responds with generic advice and no source links, Web Search may not be enabled. Go back to the Knowledge section on the Overview page and confirm the toggle is on.

If the agent doesn't respond at all, try refreshing the page and opening the Test panel again.

help me get web search working in my Copilot Studio agent

Your Content Idea Scout is alive and pulling real ideas from the web. Next up, you'll connect it to your Excel file so those ideas are captured automatically instead of disappearing after each conversation.

Build the Agent Flow to Save Ideas to Excel

Your Content Idea Scout can search the web and surface great content ideas. Right now, those ideas only exist inside the chat window. The moment you close the conversation, everything disappears.

In this step, you will build an agent flow that automatically captures each idea the agent finds and writes it to your Excel raw file. This turns a throwaway chat response into a permanent record that feeds the rest of your pipeline.

In this step, get ready to:

Create an agent flow with input parameters for idea data.
Add an Excel write action that saves ideas to your raw file.
Connect the flow to your agent and test the full pipeline.

Create a new agent flow

An agent flow is a workflow your agent can trigger during a conversation. It connects Copilot Studio to external services like Excel, email, and databases.

In Copilot Studio, select Flows in the left navigation pane.
Select New flow, then select Agent flow.

The flow designer opens with a starter template. It already includes the When an agent calls the flow trigger at the top and a Respond to the agent action at the bottom.

Select the When an agent calls the flow trigger card at the top of the canvas.
Select Add an input and choose Text as the type.
Name the first input Idea.
Add a second Text input named SourceURL.
Add a third Text input named Topic.
Add a fourth Text input named Format.

You should now see the trigger card displaying four input parameters: Idea, SourceURL, Topic, and Format.

What is an agent flow?

Agent flows are low-code automations you build inside Copilot Studio. When added as a tool, the agent's orchestrator can call the flow at runtime to retrieve data or perform actions on behalf of the user.

The When an agent calls the flow trigger makes the flow callable from your agent. The Respond to the agent action sends a reply back to the chat when the flow finishes.

Add the Excel write action

With the trigger configured, you need to tell the flow what to do with the incoming data. You will connect it to the Excel Online (Business) connector so each idea is written directly to your raw file.

Below the trigger card, select the + button to add an action.
Search for Excel Online (Business) in the connector search bar.
Select the Add a row into a table action.

Why Excel Online (Business)?

This connector works with Excel files stored in OneDrive for Business or SharePoint. It requires your data to be formatted as a Table (which you set up in Step 1) so it can identify the correct columns to write to.

In the Location dropdown, select your OneDrive.
In the Document Library dropdown, select OneDrive.
In the File field, browse to and select Content-Ideas-Raw.xlsx.
In the Table dropdown, select the table you created earlier.

Now map the columns so the flow knows which input goes where.

In the Idea field, select the Idea input from dynamic content.
In the Source URL field, select the SourceURL input from dynamic content.
In the Topic field, select the Topic input from dynamic content.
In the Format field, select the Format input from dynamic content.
In the Date Found field, switch to the expression tab and type utcNow().

What does utcNow() do?

The utcNow() expression is a built-in function that returns the current UTC timestamp when the flow runs. Every idea gets an automatic date stamp without you or the agent needing to provide one.

Select the Respond to the agent action at the bottom of the flow.
Add an output message such as Idea saved to the raw file successfully. so the agent can confirm the save to the user.
Select Publish in the top-right corner to save and publish the flow.

After publishing, you should see a confirmation message indicating your flow is live.

Testing does not consume credits

Testing an agent flow from the test chat does not consume Copilot Studio capacity. You can test as many times as you need without worrying about billing.

🙋‍♀️ Don't see your Excel file in the file picker?

Make sure Content-Ideas-Raw.xlsx is saved directly in your OneDrive root or in a folder you can navigate to. If the file was created elsewhere, move it to your OneDrive.

Also confirm the data is formatted as a Table (with filter arrows on the headers). The connector cannot detect plain cells with headers.

Why can't the flow find my Excel table?

Connect the flow to your agent and test

Your flow is published and ready. Now you need to register it as a tool on your Content Idea Scout agent so the agent knows the flow exists and when to call it.

In Copilot Studio, select Agents in the left pane.
Select your Content Idea Scout agent.
Go to the Tools page.
Select Add a tool.
Select Flow to show the list of available published flows.
Select the flow you just created, then select Add and configure.
In the Description field, enter Save a content idea to the raw Excel file for later review. so the agent knows when to use this tool.
Select Save.

Why does the description matter?

The agent's orchestrator reads the tool description to decide when to call the flow. A clear, specific description helps the agent understand that this flow should be triggered whenever a user wants to save an idea to the raw file.

Time to test the full pipeline end-to-end. Before you run this test, what do you think will happen in your Excel file when the agent saves an idea?

Open the Test your agent panel in the bottom-right corner.
Type the following message: Find me a content idea about remote work productivity and save it to the raw file.
Wait for the agent to respond. It should search the web, find an idea, and then confirm it saved the idea to the raw file.
Go to OneDrive and open Content-Ideas-Raw.xlsx.

You should see a new row in your table with the idea, source URL, topic, and a date stamp. This confirms your entire pipeline is working from chat to spreadsheet.

No new row in your Excel file?

Check these common causes:

Make sure the flow is published (not just saved). Return to the Flows page and confirm its status shows as active.
Confirm the flow is connected to the agent as a tool. Check the Tools page on your agent and verify the flow appears in the list.
If the agent responded with ideas but did not mention saving, update your agent's instructions to include guidance like: "When a user wants to save an idea, use the save tool."

Help me troubleshoot why my agent flow is not writing to Excel.

Your agent can now find ideas and save them permanently. Next up, you will add a one-click approval flow so only the best ideas make it into your structured masterfile.

Add the Human Approval Gate

Your agent can now find content ideas and save them to the raw file automatically. But every idea lands there with no quality filter. Great ideas sit alongside irrelevant ones, and there is no way to tell which should move forward.

You need a human checkpoint. In this step, you will create an automated flow that watches for changes in your raw file. When you type Approved in an idea's Status column, the flow automatically copies that idea into the masterfile with an auto-generated Content ID. Your approval is just typing one word.

In this step, get ready to:

Create an automated flow triggered when a row changes in Excel.
Add a condition that checks for approved status and prevents duplicates.
Test the full pipeline by approving an idea from the raw file.

Create the automated flow

Instead of hunting for a button in Excel's ribbon, you will use a trigger that fires automatically whenever a row changes. The flow reads the updated row, checks if you typed Approved in the Status column, and handles the rest.

In Copilot Studio, select Flows in the left navigation.
Select New flow, then select Automated cloud flow.
Name the flow Approve Content Idea.
Under Choose your flow's trigger, search for Excel Online (Business).
Select the When a row is modified trigger.
Select Create.

The flow designer opens with the trigger card at the top. Configure it to point at your raw file.

In the Location dropdown, select your OneDrive.
In the Document Library dropdown, select OneDrive.
In the File field, browse to and select Content-Ideas-Raw.xlsx.
In the Table dropdown, select the table you created earlier.

What is the When a row is modified trigger?

This trigger fires every time any cell in a table row changes. The flow receives the entire updated row as dynamic content, so you can read any column value without extra actions.

Because it fires on every edit (not just Status changes), you need a condition inside the flow to filter for only the changes you care about.

Add the approval condition and actions

The trigger fires on any row change, so the flow needs to check two things: did you set Status to Approved, and has this row already been processed? The Date Approved column acts as the guard. If it already has a value, the flow skips that row.

Below the trigger, select + to add an action.
Search for Condition and select it.

You need two conditions combined. Set the first row of the condition:

Set the left value to the Status column from dynamic content.
Set the operator to is equal to.
Set the right value to Approved.

Now add the second row to guard against duplicates:

Select + Add to add another row to the condition.
Set the left value to the Date Approved column from dynamic content.
Set the operator to is equal to.
Leave the right value empty (blank).

Why check both Status and Date Approved?

The Status check ensures the flow only acts when you explicitly mark an idea as approved. The Date Approved check prevents the flow from processing the same row twice. Once the flow writes a timestamp into Date Approved, this condition fails on any future trigger for that row.

Together they mean: only process a row that you just approved AND that has not been processed before.

Now fill in the If yes branch (Status is Approved AND Date Approved is blank). This is where the actual approval happens.

In the If yes branch, add an Initialize variable action.
Set Name to ContentID.
Set Type to String.
Set Value to the expression:

concat('PID-', formatDateTime(utcNow(), 'yyMMdd'))

What does this expression do?

The concat() function joins the literal prefix PID- with today's date formatted as YYMMDD. For example, an idea approved on 5 August 2026 gets the Content ID PID-260805.

Below the variable, add an Add a row into a table action (Excel Online Business).
Set Location to your OneDrive.
Set Document Library to OneDrive.
Set File to Content-Masterfile.xlsx.
Select your table, then map the columns:

Content ID = the ContentID variable.
Idea = Idea column from the trigger.
Format = Format column from the trigger.
Source URL = Source URL column from the trigger.
Topic = Topic column from the trigger.
Status = type the literal text Approved.
Date Approved = use the expression utcNow().

Now stamp the raw row so the guard catches any future triggers on this same row.

Below the Add a row action, add an Update a row action (Excel Online Business).
Set Location to your OneDrive.
Set Document Library to OneDrive.
Set File to Content-Ideas-Raw.xlsx.
Select your table.
In the Key Column dropdown, select Idea.
In the Key Value field, select the Idea column from the trigger's dynamic content.
Map the Content ID column to the ContentID variable.
Map the Date Approved column to the expression utcNow().

Leave the If no branch empty. If the condition fails (Status is not Approved, or Date Approved is already filled), the flow simply ends without doing anything.

Select Save at the top of the designer.

You should see a green confirmation that the flow saved successfully.

Flow not saving?

Check that every action has its required fields filled in. The most common miss is the Table dropdown in the trigger or the Add a row action. If a field shows a red border, it needs a value.

Also confirm the expression fields use the Expression tab (not the Dynamic content tab) when entering utcNow() and concat().

Help me troubleshoot why my approval flow won't save.

Test the full approval pipeline

Time to run the flow for real. Before you do, predict what you expect to happen: you type Approved in a row's Status column, the flow detects the change, and the masterfile should receive a new entry with an auto-generated Content ID.

Open Content-Ideas-Raw.xlsx in Excel Online (on OneDrive).
Find a row that has idea data (from your earlier test in Step 3).
In that row's Status column, type Approved and press Enter.
Wait 1-2 minutes for the flow to trigger and complete.

Why the wait?

The When a row is modified trigger polls for changes on a schedule (typically every 1-3 minutes). Unlike a button flow that runs instantly, automated triggers check for changes periodically. Your flow will run shortly after the polling interval detects your edit.

Now verify the results.

Check the row you just approved in Content-Ideas-Raw.xlsx. The Content ID column should show a value like PID-260805 and Date Approved should show a UTC timestamp.
Open Content-Masterfile.xlsx in OneDrive.
Confirm the approved idea appears as a new row with Content ID, Idea, Format, Source URL, Topic, Status Approved, and Date Approved filled in.

What should I see in the masterfile?

Your Content-Masterfile.xlsx table should now have one row. The Content ID column shows something like PID-260805, Status shows Approved, and the Publish Date column is empty (you fill that in when you schedule the content).

Now test the duplicate guard. Edit the same row again (change any other cell) and wait for the flow to trigger.

In the same row you just approved, edit the Topic column (add a space or change a word).
Wait 1-2 minutes for the flow to trigger again.
Open Content-Masterfile.xlsx and confirm it still has only one row for that idea.

Seeing duplicate rows in the masterfile?

Check your condition has two rows: Status is equal to Approved AND Date Approved is equal to blank. If the Date Approved check is missing, the flow processes the row every time any cell changes.

Also confirm the Update a row action in the If yes branch is writing utcNow() to the Date Approved column. If this stays empty after the first run, the guard has nothing to check.

Help me fix my duplicate approval guard.

Your full content idea pipeline is live. The agent finds ideas, saves them to the raw file, and you approve the best ones by typing one word in the Status column. The flow handles the rest automatically.

Secret mission

Add AI Pre-Scoring

Your pipeline works, but the human reviewer sees every idea equally. What if the AI could rank ideas by relevance before they reach you? Add a scoring rubric so the best ideas float to the top of your approval queue.

Clean Up Your Resources

Decide whether to keep your resources running, pause them to come back later, or delete them entirely. This project runs on a free Copilot Studio trial with Excel files stored on OneDrive, so there are no ongoing costs.

Resources you used:

Content Idea Scout agent in Copilot Studio (free trial, no publish costs).
Two agent flows in Copilot Studio: Save to Raw File and Human Approval (testing does not consume Copilot Studio capacity).
Content-Ideas-Raw.xlsx on OneDrive (no cost).
Content-Masterfile.xlsx on OneDrive (no cost).

Keep everything running

No action needed. Choose this if you plan to continue building on this agent for subsequent projects like script generation or a full video pipeline.

Your Copilot Studio trial stays active for 30 days. After the trial expires, the agent continues to work for up to 90 days.
Your Excel files remain on OneDrive indefinitely at no cost.
Your agent flows remain saved and published, ready for future use.

Pause - I'll come back to this later

Shut down running processes to free up resources while keeping your files intact for later.

Do nothing. The trial auto-extends for 90 days after expiry, and your Excel files remain on OneDrive indefinitely.
Your agent and flows stay in place and can be resumed at any time within the trial window.

Delete - I don't want to use this again

Remove all project resources and start fresh. Follow these steps to clean up everything.

Delete the agent: In Copilot Studio, go to the Agents page. Select Content Idea Scout. On the top menu bar, select the three dots (...) next to Settings, then select Delete. Enter the agent's name to confirm and select Delete agent. This permanently deletes the agent and all associated data after a few minutes.
Delete the agent flows: In Copilot Studio, go to Flows in the left navigation. Select your Save to Raw File flow, go to the Overview tab, and select Delete flow. Repeat for the Human Approval flow.
Delete the Excel files: Go to OneDrive (onedrive.com). Select Content-Ideas-Raw.xlsx by clicking the circle checkbox next to it. Then also select Content-Masterfile.xlsx. Click Delete on the toolbar at the top of the page.
The Copilot Studio trial environment will expire on its own after 30 days. No manual action is required to close it.

Nice Work!

You just built a complete AI content idea pipeline in Copilot Studio. From a single chat prompt, your agent searches the live web, captures ideas in a raw file, and you approve the best ones with a single click directly in Excel.

Here's what you learned:

Built an AI agent in Copilot Studio that searches the live web for content ideas on any topic and writes them to an Excel raw file using agent flows.
Implemented a human-in-the-loop approval gate using a button flow that lets you approve ideas directly in Excel with auto-generated Content IDs and duplicate protection.
Created a structured masterfile that only receives approved ideas with Content ID, format tags, and timestamps filled automatically.
Secret Mission: Added AI pre-scoring with a relevance rubric so you can sort and prioritize ideas before approving them.

Ready to quiz yourself?