Find new business alerts from Instagram and news sites

Monitor Instagram posts and news articles to discover new business openings, expansions, funding announcements, and other signals. PromptLoop's signals task automatically searches, crawls content, and extracts structured information on a recurring schedule.

Prerequisites#

• PromptLoop account with task creation access
• A search query that surfaces relevant Instagram posts or news articles (e.g., "new yoga studio opening Austin")
• Basic understanding of search operators (optional but helpful)

Steps#

1. Navigate to Task Template Library in PromptLoop and click Templates.

2. Select Signals Task from the task list. This uses the time-filtered search function optimized for monitoring recent content.

3. Define your search query in the task configuration. Your query should target the type of business alert you want to find:

   • "new restaurant opening [city name]"
   • "hair salon grand opening [location]"
   • "tech startup raises funding [industry]"
   • "[company type] expansion announcement"

4. Set the time period to control how far back to search. Options include:

   • Past week
   • Past month

5. Configure what data to extract. The default signals task returns:

   • Company Name
   • Company Website
   • Description of Event or Company

   You can customize to add fields like:

   • Location or address
   • Announcement date
   • Funding amount (for fundraising alerts)
   • Contact information
   • Social media links

6. Add filters (optional) to refine results. Filter by:

   • Location keywords in description
   • Minimum content length
   • Presence of specific terms (e.g., "opening", "launched")

7. Enable recurring schedule. Toggle "Run on Schedule" and select your frequency:

   • Daily
   • Weekly (recommended for most monitoring)
   • Bi-weekly
   • Monthly

8. Test with a single search first. Run the task once to verify it returns the expected results before scheduling.

9. Create a dataset to collect results. Either:

   • Create a new empty dataset that the scheduled task will populate
   • Use an existing dataset to append new findings

10. Activate the schedule and monitor the first few runs to ensure quality.

Example#

You want to track new hair salon openings in Austin to offer business services.

Task Configuration:

• Search query: "new hair salon opening Austin Texas"
• Time period: Past week
• Schedule: Every Monday at 9:00 AM
• Extracted fields:
  • Salon Name
  • Website
  • Opening Description
  • Instagram Link
  • Address

Sample Results:

The task runs every week and adds new findings to your dataset automatically.

Understanding signals task crawling#

The signals task works differently from standard research tasks:

What it does:

1. Searches Google/Bing for content matching your query within the time period
2. Filters results to recent posts, articles, and announcements
3. Crawls each relevant link to extract full content
4. For Instagram and social media links, uses search engine snapshots when direct crawling is limited
5. Extracts your specified data points using AI
6. Adds results to your dataset

Instagram handling:

• Signals tasks can detect Instagram URLs in search results
• Content is extracted from search engine previews and cached pages
• Works best for public business accounts and posts
• May have limitations on private accounts or stories

News site handling:

• Full article crawling for most news sites
• Extracts content even from sites with paywalls (using preview content)
• Handles dynamic loading and JavaScript rendering

Quality checks#

After your first scheduled run:

• Review the dataset for false positives (irrelevant results)
• Check if any expected businesses were missed
• Verify that extracted fields have accurate data
• Adjust filters to improve precision

If results are too broad:

• Add more specific location terms to your query
• Use filters to require certain keywords in descriptions
• Narrow the time period to reduce volume

If results are too narrow:

• Broaden your search query (e.g., "salon" instead of "hair salon")
• Expand the time period
• Remove restrictive filters

Tips#

• Start with weekly schedules rather than daily to avoid overwhelming results
• Use multiple tasks for different locations or business types
• Combine search operators like "new opening" AND "announced" for precision
• Instagram is best for consumer-facing businesses (retail, restaurants, salons)
• News sites work better for B2B companies, tech startups, and larger announcements
• Export results to HubSpot or Salesforce to trigger follow-up workflows
• Set up email alerts when new businesses are found (via dataset webhook)

Advanced search techniques#

Improve your query with search operators:

Location targeting:

• "new hair salon" site:instagram.com Austin
• beauty salon opening "San Francisco"

Time-based keywords:

• "now open" OR "grand opening" OR "just launched"
• announced funding OR raised capital

Exclude irrelevant results:

• new hair salon opening -job -hiring (excludes job postings)
• salon opening -permanently (excludes closure news)

Source targeting:

• site:instagram.com for Instagram-only
• site:techcrunch.com OR site:venturebeat.com for tech news

FAQs#

Can signals tasks track Instagram stories?
No. Signals tasks work with publicly indexed Instagram posts that appear in search engines. Stories expire and are not typically indexed.

How far back can I search?
You can search up to 1 year back, but results quality decreases for older content. Best results come from the past week to past month.

Will this capture every single business opening?
No. Signals tasks find businesses that have public announcements in search-indexed content. Some businesses may open without significant online presence or announcement.

Can I monitor competitor announcements?
Yes. Use queries like "[competitor name] announces" or "site:competitor.com/blog" to track specific company news.

What happens if a scheduled run finds no results?
The task completes successfully but adds no rows to your dataset. This is normal if there are no new announcements during that period.

How many results will each run generate?
Varies significantly based on your query. A broad query in a major city might return 10-50 results per week, while a niche query might return 0-5.

Related#

• Create your first PromptLoop task
• Monitor autoloop jobs
• Build targeted business lists with AI