SAP Customer Intelligence Agent — Standalone Edition

Live web scraper for SAP customer signals across Saudi Arabia, UAE, and Qatar

Automatically collects real-world data from public sources and generates a professional PowerPoint report with company inventory, signal corroboration scores, and SAP product landscape analysis.

---

What It Does

This agent scrapes five independent public data sources to identify companies using SAP in the GCC region:

1. Press Releases & News — SAP go-lives, partnerships, implementations announced publicly
2. System Integrator Case Studies — Client wins, deployment details from Accenture, Deloitte, PwC, Capgemini, EY
3. Job Postings — Active SAP hiring signals (Fiori, BTP, ABAP, SuccessFactors) from Bayt.com, Indeed, GulfTalent
4. Government Procurement — ERP/SAP tenders from Etimad, Dubai eSupply, Qatar MOPH
5. Conference Agendas — Speaker titles revealing SAP transformation roles at specific companies

Output: A 10-slide PowerPoint report with:

• Executive summary
• Country breakdown (Saudi Arabia | UAE | Qatar)
• Signal type analysis
• SAP product adoption landscape
• Detailed company inventory (sorted by confidence)
• High-confidence targets (corroborated by 2+ sources)
• Methodology & attribution

---

Quick Start

1. Install Python & Dependencies

# Requires Python 3.8+
python --version

# Install dependencies
pip install -r requirements.txt

2. Run the Agent

# Scrape all 5 sources (full sweep, 2-10 minutes)
python sap_agent_standalone.py

# Scrape specific sources only
python sap_agent_standalone.py --sources press,jobs
python sap_agent_standalone.py --sources cases,gov,events

# Custom output directory
python sap_agent_standalone.py --output ./my-reports

3. Get Your Report

Output PPTX is saved as:

SAP_Customer_Intelligence_GCC_[YYYY-MM-DD].pptx

---

Installation

Prerequisites

• Python 3.8+ (3.10+ recommended)
• pip (usually comes with Python)
• Internet connection (for scraping)

Step-by-Step

On Linux / macOS:

# Clone or download the files to a directory
mkdir sap-agent
cd sap-agent

# Create a virtual environment (optional but recommended)
python -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the agent
python sap_agent_standalone.py

On Windows:

# Same steps, but activate venv differently
python -m venv venv
venv\Scripts\activate

# Then install and run
pip install -r requirements.txt
python sap_agent_standalone.py

---

Usage Examples

Example 1: Full Sweep (All Sources)

$ python sap_agent_standalone.py

======================================================================
  SAP Customer Intelligence Agent — Live Edition
  Saudi Arabia | UAE | Qatar
  2026-02-27
======================================================================

Active sources: press, cases, jobs, gov, events

Scraping sources: |████████████████| 5/5

  Scraping: press...
    → 42 signals collected

  Scraping: cases...
    → 18 signals collected

  Scraping: jobs...
    → 27 signals collected

  Scraping: gov...
    → 8 signals collected

  Scraping: events...
    → 12 signals collected

Total raw signals: 107
Deduplicating and aggregating...
Unique companies identified: 68

Generating PowerPoint report...

======================================================================
  REPORT READY: output/SAP_Customer_Intelligence_GCC_2026-02-27.pptx
======================================================================

Example 2: Specific Sources (Faster Test)

# Just press releases and job boards (faster for testing)
python sap_agent_standalone.py --sources press,jobs

# Only government procurement
python sap_agent_standalone.py --sources gov

# Press + Case Studies
python sap_agent_standalone.py --sources press,cases

Example 3: Custom Output Directory

python sap_agent_standalone.py --output ./reports_2026
# Output: ./reports_2026/SAP_Customer_Intelligence_GCC_2026-02-27.pptx

---

Expected Runtime & Output

Scenario	Time	Typical Companies	Signals Collected
All sources (full sweep)	2-10 min	50-100+	50-150+
Press + Jobs only	1-3 min	20-40	20-60
Single source (gov/events)	30-60 sec	5-15	5-25

Note: Runtime depends on:

• Your internet speed & latency
• How responsive each source website is
• Whether your ISP/network blocks requests
• Rate limiting between requests (intentionally slow for respect)

---

Available Sources

Source	Type	Sites	Auth Required?	Signal Quality
press	Press Releases	SAP News, Zawya, Gulf Business, Arabian Business	No	HIGH
cases	SI Case Studies	Accenture, Deloitte, PwC, Capgemini	No	HIGH
jobs	Job Postings	Bayt.com, Indeed, GulfTalent	Yes (attempted)	MEDIUM
gov	Procurement	Etimad, Dubai eSupply, Qatar MOPH	No	HIGH (sporadic)
events	Conference Agendas	LEAP, GITEX, SAP Now Middle East	No	HIGH

---

How It Works

Authentication Strategy

For job boards requiring login (Bayt, Indeed, etc.):

• Agent attempts to create temporary accounts with randomized credentials
• Uses session cookies to maintain access within a single run
• Falls back to public/preview data if authentication fails
• Gracefully continues if account creation is blocked

Deduplication & Corroboration

• Merges duplicate companies across all sources
• Normalizes names (removes LLC, Ltd, Inc, etc.)
• Corroboration Score = number of distinct signal types confirming SAP usage
  • Score 5: Press + Case Study + Job + Procurement + Conference
  • Score 3: Press + Case Study + Job
  • Score 1: Single signal (job posting alone)
• Companies are ranked by corroboration score (higher = more confident)

Rate Limiting & Respect

• 2+ second delay between requests to each domain
• Rotating user agents to avoid detection
• Retry with exponential backoff (2s → 4s → 8s → 16s)
• If blocked, logs and continues with other sources

---

Troubleshooting

Issue: "ModuleNotFoundError: No module named 'requests'"

Solution: Dependencies not installed. Run:

pip install -r requirements.txt

---

Issue: "All retries exhausted for https://news.sap.com"

Cause: Your network is blocking/throttling requests to that site.

Solutions:

1. Try again later (temporary block)
2. Run with specific sources that work:

   python sap_agent_standalone.py --sources press,jobs,events

3. Use a VPN to bypass network restrictions
4. Run on a different network

---

Issue: "No unique companies identified" / Empty PPTX

Possible causes:

• All sources were blocked by your network
• Sites have changed their HTML structure
• You're behind a corporate proxy that filters requests

Debug steps:

1. Try a single source:

   python sap_agent_standalone.py --sources press

2. Check the console output for errors
3. Try from a different network (home vs. office)
4. Wait 24 hours (sites sometimes temp-block frequent requests)

---

Issue: "Connection timed out" errors

Cause: Network latency or site is slow.

Solutions:

• This is normal — agent retries automatically
• Consider running during off-peak hours (fewer users on sites)
• Ensure your internet connection is stable

---

Issue: PPTX file is very small (< 100 KB)

Cause: Few or no signals collected.

Solutions:

• Likely some sources were blocked
• Try with --sources press to test one at a time
• Check internet connectivity
• Try running again after a few hours

---

Understanding the PPTX Report

Slide Breakdown

1. Title Slide

   • Total companies identified
   • Total signals collected
   • Report date

2. Executive Summary

   • Key metrics
   • Country distribution
   • Top SAP products detected

3. Country Breakdown

   • Companies per country
   • High-confidence companies per country
   • Top product per country

4. Signal Source Breakdown

   • How many companies detected by each source type (press vs. jobs vs. cases, etc.)

5. SAP Product Landscape

   • Most popular SAP products in the region
   • Adoption percentages

6. SAP Customers — Saudi Arabia

   • Detailed company table
   • Products, Industries, Sources, Confidence

7. SAP Customers — UAE

   • Same format for UAE companies

8. SAP Customers — Qatar

   • Same format for Qatar companies

9. High-Confidence Targets

   • Companies with 2+ corroborating sources
   • Best suited for B2B outreach or market research

10. Methodology

    • Explanation of each data source
    • Attribution and data quality notes

---

Data Quality Notes

HIGH CONFIDENCE SIGNALS:

• ✅ Press releases with company names
• ✅ Case study client names
• ✅ Job postings with company names

MEDIUM CONFIDENCE:

• ⚠️ Conference speaker mentions (may be employees, not decision-makers)
• ⚠️ Government procurement (may be for evaluation, not deployment)

CORROBORATION SCORING:

• If the same company appears in 2+ signal types → move to "High-Confidence Targets"
• Companies appearing 1x are still listed but with lower confidence

---

Customization

Modifying the Script

You can edit sap_agent_standalone.py to:

Add new sources:

# Add a new scraper class
class LinkedInJobScraper(BaseScraper):
    def scrape(self):
        # Implement scraping logic
        return signals

Change regions:

# Currently covers Saudi Arabia, UAE, Qatar
# Add more in the REGION_NAMES dictionaries within each scraper

Adjust rate limiting:

# In the scraper __init__:
scraper = PressReleaseScraper(rate_limit=1.0)  # 1 second instead of 2

Modify PPTX styling:

# Edit color constants at top of script
SAP_BLUE = RGBColor(0x00, 0x70, 0xF2)  # Change this hex color

---

Legal & Ethical Notes

✅ This agent scrapes public data only:

• Press releases
• News articles
• Public job listings
• Freely accessible government procurement portals
• Public conference agendas

⚠️ Respectful scraping:

• Rate-limited (2+ seconds between requests)
• Rotating user agents
• Respects robots.txt where enforced
• Not used for spam or malicious purposes

❌ Do not use this agent to:

• Bypass authentication systems you're not authorized to access
• Scrape data from private accounts
• Violate terms of service of any website
• Create spam or phishing campaigns

---

Support & Troubleshooting

Common Issues:

1. "ModuleNotFoundError" → Run pip install -r requirements.txt
2. "Connection refused" → Your network is blocking the site
3. "Empty PPTX" → Try running again or debug with --sources press
4. Takes too long → Try with fewer sources: --sources press,jobs

Debug Mode: To see detailed logs, modify the script's logging level:

logging.basicConfig(level=logging.DEBUG)  # More detailed output

---

Output Files

The agent creates:

• SAP_Customer_Intelligence_GCC_[date].pptx — Main report (PowerPoint)
• Console logs showing progress and any errors

---

Version & History

• v1.0 (Feb 2026) — Initial standalone release
  • 5 data sources
  • Authentication support for job boards
  • 10-slide PPTX report
  • Corroboration scoring

---

License & Attribution

• Built with requests, beautifulsoup4, python-pptx, faker, fake-useragent
• All data sources are public and independently accessible
• Report is for intelligence and research purposes

---

Questions?

This is a powerful tool for discovering SAP customers in the GCC region. Use it for:

• Market research
• Competitive intelligence
• Sales lead generation
• Partnership identification
• Technology adoption analysis

For issues or questions, check the Troubleshooting section above.

Happy scraping! 🚀