SocialScore — Customer Intelligence Layer

2.1 The Two Layers #

SocialScore is deliberately split into two distinct layers. Understanding this helps you choose the right access mode for your team:

A fintech risk team cares most about Layer 1 — data depth and fraud signals. A CMO cares most about Layer 2 — segmentation and audience activation. Both use the same platform.

4.1 Email Signals #

Returned in email_details. Core identity and quality signals derived from the email address.

Domain Details (email_domain_details)

4.2 Phone Signals #

Returned in phone_details. Phone signals verify the number and reveal messaging platform presence.

4.3 Profile & Account Signals #

When platform data is available (primarily via LinkedIn and other public profile networks), the following account-level details are returned:

4.4 Interest Categories #

Returned in general_info.user_interests. Each category returns a score (0–1000) and a difference_to_average_score — the percentage difference vs. the average user in the same country. Local benchmarking makes the signal directly useful for your market.

How to read the score: 500 = country average. 800 = top 30% in this category. 200 = bottom 20%. A score of 923 in Education means this person is in the top 8% of all users in your country for education-related platform activity.

4.5 Health Scores — The Five Quality Indicators #

The Health Score is not a single number — it is five complementary scores that together answer: "How healthy is this identity, and how much should I trust it?" This is the primary audience quality tool for marketers and risk teams alike.

4.6 Predictive Signals — Segment Matching #

Returned in similarity_scores array. This is SocialScore's predictive layer — it tells you how similar a new, unknown identity is to your best existing customers.

Multiple segments return multiple entries in one API call. This means a single request can simultaneously tell you: "this person matches your High LTV Fashion segment at 84%, matches your Fraud Risk segment at 12%, and matches your VIP Loyalty segment at 71%." One call, three decisions.

4.7 Custom Screening Lists #

SocialScore supports name-based screening against structured reference lists as part of the enrichment pipeline. This is an optional module that operates independently from the identity enrichment signals. Screening is performed against structured name lists — not behavioral scoring — and returns a match indicator, not a decision.

4.8 Behavior Pixel — Real-Time On-Site Signals In Development

The SocialScore Pixel is a lightweight JavaScript tag (7-line async embed, non-blocking) that merchants deploy once on their storefront. It collects behavioral and hardware signals from every site visit — enriching the Customer Intelligence Layer with real-time session context that OSINT signals alone cannot provide.

When a visitor converts (submits email/phone at checkout or registration), Pixel signals are instantly merged with their identity enrichment profile — giving you a richer Day 1 persona than email-only enrichment can produce. The Pixel also enables SocialScore to begin building a preliminary behavioral profile for anonymous visitors before they identify themselves.

Behavioral Signals

Device & Hardware Signals

Attribution & Context Signals

Traffic Classification Output

Every session is classified in real time into one of four categories — enabling automated routing before any human review:

Classification uses a composite bot_score (0–100) computed server-side from behavioral anomalies, hardware signals, and WebDriver detection. No ML inference latency — result is available within the same session response.

5.1 Dashboard #

The SocialScore dashboard is fully customizable to your business sector. It provides rich predefined reports and filters across five primary views:

• Digital Footprint — which social platforms your audience uses and at what penetration
• Digital Communication Preference — which messaging channels (WhatsApp, Viber, Telegram) your audience actually has active
• Personal Interests — interest scores ranked from most to least interested, benchmarked against local average
• Health Scores — email, phone, social, and fraud scores across your audience
• Demographics & Contact Validation — location distribution, language, contact quality

Each section includes AI Hints — data-driven recommendations generated automatically from your audience's signals (e.g. "61% of your audience is on LinkedIn — prioritize this channel for B2B outreach").

Navigation: Dashboard (audience overview) · Manual Lookup (single or bulk check) · Transactions (full history, segmentation, export)

Dashboard — audience segment view showing Digital Footprint, Digital Communication Preference, Personal Interests, Health Scores, and Demographics for a selected audience label.

5.2 Single Profile Check #

One email + phone → full profile in under 5 seconds. Available via API or dashboard manual lookup.

When to use: Real-time enrichment at checkout, onboarding, or form submission. One profile at a time.

Manual Lookup — single user check. Enter email (required) + phone (recommended) and select country.

Single Profile result — full customer intelligence profile: General Info, LinkedIn data, Email Details, Phone Details, User Interests, Health Scores, and photos from connected platforms.

5.3 Bulk Profile Check #

Upload a CSV or send a batch API request → enriched output for your entire list. Used to clean and enrich existing databases at scale.

When to use: Database hygiene before a campaign. Scoring a full applicant pipeline. Enriching a CRM export before re-engagement. Auditing contact quality after a data acquisition.

Transaction History — full log of all profile checks (manual and bulk). Filter by date, type, label. Export to CSV. Each row links to the full profile result.

5.4 Segments — Create & Store #

Select a set of enriched profiles → define criteria → save as a named Segment. A Segment is stored as a signal profile (interest distribution + health score thresholds + platform presence pattern). Segments are reusable objects — the foundation of Predictive Matching.

What you can segment on: Interest categories · Health score ranges · Fraud score thresholds · Platform presence (e.g. "WhatsApp active") · Combinations of the above

5.5 Predictive Matching #

The core differentiator. Once a Segment is defined, SocialScore automatically scores every new incoming identity against it in real time. Every API call returns a similarity_scores array with a match percentage per segment.

• Real-time personalization at the moment of first contact — before any internal data exists on this customer
• Automatic routing of new users into pre-built buyer personas
• Dynamic pricing, offer, or content decisions based on segment membership
• For fintech: live application routing to the right risk model based on digital profile type
• For gaming: flag potential high-value vs high-risk players at registration, before first deposit

5.6 Audience Explorer #

Load a set of identifiers → explore the aggregated audience picture. Interest distribution, Health Score spread, fraud risk share, platform coverage, communication channel availability. No mandatory storage.

When to use: Auditing an existing customer base. Evaluating a new channel's traffic quality. Presenting audience intelligence to clients. Understanding who responded to a campaign before the next one.

Audience Explorer — upload or select a list, view aggregated signals, apply filters, and save as a named Segment with one click. The "Create Segment" button appears once you have filtered and reviewed the audience.

5.7 Audience Comparison #

Place two audiences side-by-side → see differences across all signals: interest profiles, Health Score distributions, fraud exposure, platform presence, communication channel split.

But the most powerful comparison is not audience vs. audience — it is your audience vs. the country average. SocialScore benchmarks every signal against the local population, so you can see exactly where your audience over-indexes and under-indexes relative to the typical consumer in your market.

Your audience vs. country average — the marketing strategy you're missing

This comparison answers three critical business questions that no CRM or analytics platform can answer from first-party data alone:

The platform data problem — and why you need your own view

Audience Comparison — select any two named audiences (segments) to compare side by side.

Audience Comparison result — two segments side by side across all signals: platform presence, interest categories, Health Scores, and country benchmark differences.

Practical example: two acquisition channels compared

5.8 Audience Builder #

Define criteria → SocialScore surfaces a new audience slice from its database. No existing customer list needed. You describe the persona you want (interest scores, health score ranges, platform presence), the platform generates a targetable audience.

When to use: Entering a new market. Building prospecting lists. Defining a target borrower profile for a new product. Creating cold outreach audiences without purchasing data.

Audience Builder — define criteria (interests, Health Score ranges, platform presence, fraud thresholds) and create a named audience. Saved audiences become available for comparison and predictive matching.

5.9 AI Hints #

Every section in the dashboard includes AI-generated recommendations based on your audience's signals. These are not generic tips — they are computed from your actual data:

• "Your audience's Email Score average is 91/100. Prioritize email campaigns — deliverability will be high."
• "2% Fraud Score — well below average. Your acquisition channels are delivering quality traffic."
• "61% LinkedIn registered. Consider LinkedIn as a primary B2B outreach channel for this audience."
• "Fashion interest 40% above average. Test fashion-adjacent partnerships or product recommendations."

5.10 Platform Modules — Plug & Play Integrations In Development

SocialScore Platform Modules are native integrations for major e-commerce, CRM, and marketing automation platforms. Once installed, they embed SocialScore intelligence directly into your existing platform — no API code required, no separate dashboard to check. The Customer Intelligence Layer becomes part of every workflow your team already uses.

E-commerce Platform Modules

CRM & Marketing Automation Modules

Custom Integration Module

What modules enable beyond API enrichment

With modules installed, SocialScore connects traffic intelligence (who visited and from where), customer intelligence (who they are and what they want), and transaction intelligence (what they bought and whether it's risky) — all in a single Customer Intelligence Layer that sits inside your existing platform rather than alongside it.

5.11 MCP Protocol — AI Agent Commerce #

The Model Context Protocol (MCP) is an emerging standard that allows AI systems to communicate structured context with external services. SocialScore implements MCP as the intelligence layer between AI buying agents and merchant commerce endpoints — enabling autonomous AI-to-commerce transactions that are faster, more efficient, and correctly classified from the start.

How MCP integration works

1. A merchant publishes their AI commerce configuration at /.well-known/ai-commerce.json — managed through the SocialScore dashboard
2. When the SocialScore Pixel identifies an AI Agent session (bot_score analysis + WebDriver signals), it classifies the session as "AI Agent" instead of "Human"
3. The agent is routed to a lightweight JSON commerce endpoint — no HTML rendering, no JavaScript execution, no CAPTCHA
4. The agent receives structured product data, pricing, availability, and checkout instructions in machine-readable format
5. The transaction completes via MCP — faster execution, lower server load, correct pricing logic applied automatically
6. SocialScore logs the agent transaction separately from human traffic — giving merchants full visibility into their Agentic Commerce volume

{
  "store_name": "Your Store",
  "allow_autonomous_checkout": true,
  "ai_autonomous_discount_pct": 5.0,
  "bulk_threshold_units": 10,
  "bulk_discount_pct": 10.0,
  "system_instructions": "Programmatic checkout active. Return JSON only. No HTML.",
  "mcp_endpoint": "https://yourstore.com/api/mcp/v1"
}

Why SocialScore is uniquely positioned for Agentic Commerce

6.1 E-commerce — Reduce CAC, Eliminate Fraud, Personalize at Checkout #

6.2 Fintech / Credit — Better Underwriting, Lower Default Rate #

6.3 Gaming & Gambling — Smarter Acquisition, Better Compliance #

6.4 Marketing Agency — Better Results for Every Client #

6.5 AI & Agentic Commerce — Stop the Hallucination, Start the Intelligence #

AI models hallucinate when they lack context. They recommend the wrong product, apply the wrong discount, engage with a bot as if it were a person, or offer a "first purchase" discount to a customer who matches your Premium Buyers segment at 84%. The root problem is the same in every case: the AI is making decisions without knowing who it is actually dealing with.

SocialScore is the Customer Intelligence Layer that gives your AI the context it needs to stop guessing and start being accurate. Whether you're running a chatbot, a recommendation engine, a pricing algorithm, or an autonomous checkout agent, SocialScore tells the AI — in real time, before any interaction begins — whether it is talking to a human customer, a legitimate AI buyer, or a bot that should never see your real inventory.

The problem: your AI sells socks and mortgages the same way

Right now, AI-powered tools — chatbots, recommendation engines, virtual assistants — have no idea who they are talking to. A 60-year-old engineer and a 19-year-old student both land on your checkout page. Both get the same homepage. Both get the same chatbot greeting. Both get the same product recommendations.

This is not a personalisation failure — it is a context failure. The AI is not being lazy. It genuinely does not know the difference. It has no external signal about who is on the other side of the conversation. So it defaults to the average: the same tone, the same product tier, the same offer, the same channel.

The 60-year-old engineer — LinkedIn profile, 500 connections, senior role, Business Services interest in top 10%, WhatsApp confirmed, Fraud Score 0 — is treated identically to the 19-year-old student with a disposable email, no social presence, and a Gambling interest score of 780. One is your ideal premium buyer. The other may never convert. Without SocialScore, your AI cannot tell them apart and serves both with the same experience.

The three actors SocialScore helps your AI distinguish

Use Case 1: AI Chatbot / Assistant — Context injection via MCP

With SocialScore via MCP: The moment the customer's session is identified (email from cookie, phone from account login), SocialScore returns their full profile. This is injected into the AI's system prompt before the first message:

// SocialScore profile injected into LLM system prompt
"""
Customer context (SocialScore):
- Social Score: 71/100 — digitally active, established online presence
- Email quality: 98/100 — verified, non-disposable, 14+ platform registrations
- Risk signal: CLEAN (fraud_score = 0)
- Top interests: Business Services (+74% vs country avg), Education (+85%), E-commerce (+51%)
- LinkedIn: CEO at SocialScore.io, Sofia, Bulgaria — senior professional profile
- Preferred channels: WhatsApp (confirmed), Viber (confirmed), Facebook (confirmed)
- Segment match: 'Premium_Buyers_Q1' = 84% similarity
- Communication: WhatsApp-first, formal register appropriate
"""
// AI now knows: professional, high-value, clean, WhatsApp-reachable, B2B interest
// AI can immediately personalize: tone, product tier, channel recommendation, offer level

The AI no longer hallucinates a customer persona. It has real signals. It recommends the right product tier, communicates in the right register, offers WhatsApp as the first channel, and does not apply a "new customer" discount to a profile that matches the premium buyer segment at 84%.

Use Case 2: Recommendation Engine — Real identity-based recommendations

Most recommendation engines work only from purchase history. A new visitor has no history. Without SocialScore, they get generic bestsellers or trending products — the same as everyone else. CAC stays high because you're not converting the visitor efficiently.

With SocialScore, the recommendation engine receives interest scores on the first pageview:

• Travel > 700, Sports > 650, Hobbies > 600 → immediately surface travel-ready outdoor gear, not office furniture
• Fashion > 750, Business Services > 800 → surface professional attire, premium accessories — not sportswear
• Gambling > 680 + Fraud Score = 0 → legitimate player, surface premium tier products, skip "introductory" offer
• Segment match 'High LTV Fashion' > 75% → skip discount entirely, show loyalty programme instead

The result: higher conversion on first visit, lower CAC, higher average order value — because the recommendation is built on who the customer actually is, not on the absence of data.

Use Case 3: Autonomous AI Agent (Agentic Commerce)

An AI shopping agent acting on behalf of a consumer sends a request to your store. Without SocialScore, your platform serves it a full HTML page, runs all JavaScript, triggers CAPTCHA, and likely blocks or slows the agent because it looks like unusual traffic. The agent fails to complete the purchase.

With SocialScore Pixel + MCP:

1. Pixel detects the session: WebDriver flag present, GPU renderer = headless, zero mouse movement, keyboard timing = 0ms → classified as AI Agent
2. SocialScore routes the agent to the MCP JSON endpoint automatically — no HTML, no JavaScript, no CAPTCHA
3. Agent receives: structured product catalogue, real-time pricing, availability, agent-tier discount (if configured), checkout instructions
4. Agent completes purchase via JSON API in <500ms total — faster than a human browsing session takes to load the product page
5. Transaction logged separately in SocialScore dashboard: Agentic Commerce volume, agent identity, products purchased, agent-tier revenue

7.1 Endpoint & Authentication #

7.2 Request Parameters #

7.3 Response Structure #

The API returns a JSON object with four top-level keys:

{
  "email_details": {
    "email": "customer@example.bg",
    "email_score": 100,
    "is_email_disposable": false,
    "domain": "abv.bg",
    "number_of_exposure_records": 12,
    "email_history": {
      "hits": 47,
      "customer_hits": 3,
      "first_seen": "2023-09-12T14:30:00Z",
      "last_seen": "2026-07-27T09:15:00Z"
    },
    "email_domain_details": {
      "domain": "abv.bg", "tld": ".bg",
      "registered": true, "free": true, "custom": false,
      "disposable": false, "valid_mx": true, "website_exists": true,
      "dmarc_enforced": true, "spf_strict": true,
      "suspicious_tld": false, "accept_all": false
    },
    "platforms_data": [
      {
        "platform_name": "google",
        "is_registered": true,
        "account_details": {
          "photo": "https://lh3.googleusercontent.com/...",
          "activity": {
            "reviews": 1, "ratings": 2, "photos": 0,
            "answers": 11, "edits": 0
          },
          "is_enterprise_user": false,
          "last_updated": 1775633698
        }
      },
      { "platform_name": "facebook", "is_registered": true, "account_details": null },
      { "platform_name": "spotify", "is_registered": true, "account_details": null },
      { "platform_name": "instagram", "is_registered": false, "account_details": null },
      { "platform_name": "u****e", "is_registered": true, "account_details": null },
      // ... 45+ more platforms
    ],
    "security_exposure_details": {
      "haveibeenpwned_listed": true,
      "number_of_exposure_records": 12,
      "first_exposure_record": "2016-10-04",
      "breaches": [
        { "name": "LinkedIn", "domain": "linkedin.com", "date": "2016-05-21" }
      ]
    }
  }
}

{
  "phone_details": {
    "phone_number": "+359888XXXXXX",
    "phone_score": 100,
    "phone": {
      "valid": true, "disposable": false, "type": "mobile",
      "country": "BG", "carrier": "A1 Bulgaria",
      "history": { "hits": 12, "customer_hits": 1, "first_seen": "2024-01-08", "last_seen": "2026-07-27" }
    },
    "platforms_data": [
      {
        "platform_name": "whatsapp", "is_registered": true,
        "account_details": { "photo": "[base64 data]", "about": "Hey there! I am using WhatsApp." }
      },
      { "platform_name": "facebook", "is_registered": true, "account_details": null },
      { "platform_name": "viber", "is_registered": true, "account_details": { "name": "Dimitar D.", "photo": "..." } },
      { "platform_name": "telegram", "is_registered": null, "account_details": null }
    ]
  }
}

{
  "general_info": {
    "social_score": 63,    // 0–100, median 50
    "fraud_score": 0,      // 0 = clean, higher = riskier (inverted scale)
    "profile_id": "ss_abc123xyz",
    "e_signature": true,  // Borica e-signature certificate (BG market)
    "studied_languages": ["Bulgarian", "English"],
    "account_details": {
      "name": "Dimitar Dimitrov",
      "photo": "https://...",
      "linkedin": {
        "url": "https://linkedin.com/in/dimitardimitrov",
        "name": "Dimitar Dimitrov",
        "company": "SocialScore.io",
        "title": "CEO & Co-founder",
        "location": "Sofia, Bulgaria",
        "connection_count": 500,
        "photo": "https://..."
      }
    },
    "user_interests": [
      { "category": "education_and_online_learning", "score": 923, "difference_to_average_score": 84.6 },
      { "category": "business_services", "score": 874, "difference_to_average_score": 74.8 },
      { "category": "customer_services", "score": 798, "difference_to_average_score": 59.6 },
      { "category": "e_commerce_and_shopping", "score": 756, "difference_to_average_score": 51.2 },
      { "category": "social_media", "score": 718, "difference_to_average_score": 43.6 },
      { "category": "travel_and_tourism", "score": 509, "difference_to_average_score": 1.8 },
      { "category": "hobbies_and_leisure", "score": 594, "difference_to_average_score": 18.8 },
      { "category": "sports", "score": 452, "difference_to_average_score": -9.6 },
      { "category": "it_and_technology", "score": 382, "difference_to_average_score": -23.6 },
      { "category": "fashion", "score": 397, "difference_to_average_score": -20.6 },
      { "category": "gambling", "score": 383, "difference_to_average_score": -23.4 },
      { "category": "chat", "score": 380, "difference_to_average_score": -24.0 },
      { "category": "forum", "score": 358, "difference_to_average_score": -28.4 },
      { "category": "jobs_and_career", "score": 235, "difference_to_average_score": -53.0 },
      { "category": "adult", "score": 182, "difference_to_average_score": -63.6 }
    ]
  }
}

{
  "similarity_scores": [
    {
      "execution_name": "Premium_Buyers_Q1",
      "similarity_score": 84.2  // Strong match — show premium offer
    },
    {
      "execution_name": "High_Risk_Applicants",
      "similarity_score": 8.5   // Very low match — clean profile
    },
    {
      "execution_name": "VIP_Loyalty_2025",
      "similarity_score": 71.1  // Good match — eligible for loyalty program
    }
    // One entry per defined segment — all computed in a single API call
  ]
}

const response = await fetch('https://api.socialscore.io/api/v1/transactions/run', {
  method: 'POST',
  headers: {
    'X-API-KEY': 'your-api-key',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    email: 'customer@example.com',
    phone_number: '+35988xxxxxxx',
    country: 'BG'
  })
});
const profile = await response.json();

// Real-time decision logic
const { email_details, general_info, similarity_scores } = profile;
const isClean = general_info.fraud_score === 0 && !email_details.is_email_disposable;
const premiumMatch = similarity_scores.find(s => s.execution_name === 'Premium_Buyers_Q1');
if (isClean && premiumMatch?.similarity_score > 70) {
  // Show premium offer, skip discount
} else if (!isClean) {
  // Flag for review
}

// PHP cURL example
$ch = curl_init('https://api.socialscore.io/api/v1/transactions/run');
curl_setopt_array($ch, [
  CURLOPT_POST => true,
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_HTTPHEADER => [
    'X-API-KEY: your-api-key',
    'Content-Type: application/json'
  ],
  CURLOPT_POSTFIELDS => json_encode([
    'email' => 'customer@example.com',
    'phone_number' => '+35988xxxxxxx',
    'country' => 'BG'
  ])
]);
$profile = json_decode(curl_exec($ch), true);
curl_close($ch);

8. Frequently Asked Questions #

9. Support & Resources #

Support & Connection page — API key, platform status, documentation links, and contact details. Access via the profile icon in the top-right corner.

Account Profile page — manage your SocialScore account, team access, and notification settings.

Billing page — current plan, usage statistics, invoice history, and plan upgrade options.

Ready to understand your customers? #

SocialScore is the Customer Data Gateway — the world's first Customer Intelligence Layer built for B2C businesses that want to know who is behind the email, before, during, and after every interaction.

Start with the dashboard. Add the API when you're ready to automate. Connect a module when you want zero-code intelligence inside your existing platform. Every path starts with one enrichment call.

Free demo available · €0 to try · 300+ data sources · Under 5 seconds per check · GDPR compliant
Why choose between a platform and data? With SocialScore, you get both.

10. Compliance, Security & Data Governance #

8.2 What to add to your Terms & Conditions

When you integrate SocialScore, your end-user Terms of Service and Privacy Policy should include the following disclosures. These are standard for any business using third-party data enrichment — including businesses using Google Analytics, Meta Pixel, or any CRM enrichment tool.

Recommended T&C disclosure language

Your Privacy Policy and/or Terms of Service should include a section that covers data enrichment. The following elements should be addressed (consult your legal counsel for jurisdiction-specific wording):

• Purpose: "We use third-party data enrichment services to improve service quality, including fraud prevention, contact validation, and personalisation of our communications and product recommendations."
• Legal basis (GDPR): "This processing is carried out on the basis of our legitimate interests (GDPR Art. 6(1)(f)) in preventing fraud, improving the relevance of our communications, and optimising our customer experience."
• Data categories: "We may pass your email address, phone number, and country to our enrichment partner to retrieve publicly available digital footprint signals associated with those identifiers."
• User rights: "You have the right to object to this processing (GDPR Art. 21) and to request deletion of any stored enrichment profile associated with your contact details. Contact us at [your contact email] to exercise these rights."
• No automated decisions: "Enrichment signals are used to inform human-reviewed or human-configured business processes. No automated individual decisions with legal or similarly significant effects are made solely on the basis of enrichment data."

8.3 GDPR Framework

8.3 EU AI Act Position

SocialScore enrichment signals used for marketing personalization, audience segmentation, and contact validation are not classified as high-risk AI systems under EU AI Act Annex III.

If you use SocialScore signals as inputs to systems that make automated decisions on creditworthiness, employment, insurance, or essential services eligibility, those downstream systems may themselves qualify as high-risk AI under Annex III — independently of SocialScore. Your legal and compliance team must assess your complete decision pipeline, not only SocialScore's role within it.

SocialScore does not perform profiling that produces legal or similarly significant effects on individuals within the meaning of GDPR Art. 22. All profiling outputs are enrichment signals for human-reviewed or human-configured business logic.

Ready to understand your customers? #

SocialScore is the Customer Data Gateway — the world's first Customer Intelligence Layer built for B2C businesses that want to know who is behind the email, before, during, and after every interaction.

Start with the dashboard. Add the API when you're ready to automate. Connect a module when you want zero-code intelligence inside your existing platform. Every path starts with one enrichment call.

Free demo available · €0 to try · 300+ data sources · Under 5 seconds per check · GDPR compliant
Why choose between a platform and data? With SocialScore, you get both.

8.2 What to add to your Terms & Conditions

When you integrate SocialScore, your end-user Terms of Service and Privacy Policy should include the following disclosures. These are standard for any business using third-party data enrichment — including businesses using Google Analytics, Meta Pixel, or any CRM enrichment tool.

Recommended T&C disclosure language

Your Privacy Policy and/or Terms of Service should include a section that covers data enrichment. The following elements should be addressed (consult your legal counsel for jurisdiction-specific wording):

• Purpose: "We use third-party data enrichment services to improve service quality, including fraud prevention, contact validation, and personalisation of our communications and product recommendations."
• Legal basis (GDPR): "This processing is carried out on the basis of our legitimate interests (GDPR Art. 6(1)(f)) in preventing fraud, improving the relevance of our communications, and optimising our customer experience."
• Data categories: "We may pass your email address, phone number, and country to our enrichment partner to retrieve publicly available digital footprint signals associated with those identifiers."
• User rights: "You have the right to object to this processing (GDPR Art. 21) and to request deletion of any stored enrichment profile associated with your contact details. Contact us at [your contact email] to exercise these rights."
• No automated decisions: "Enrichment signals are used to inform human-reviewed or human-configured business processes. No automated individual decisions with legal or similarly significant effects are made solely on the basis of enrichment data."

8.3 GDPR Framework

8.3 EU AI Act Position

SocialScore enrichment signals used for marketing personalization, audience segmentation, and contact validation are not classified as high-risk AI systems under EU AI Act Annex III.

If you use SocialScore signals as inputs to systems that make automated decisions on creditworthiness, employment, insurance, or essential services eligibility, those downstream systems may themselves qualify as high-risk AI under Annex III — independently of SocialScore. Your legal and compliance team must assess your complete decision pipeline, not only SocialScore's role within it.

SocialScore does not perform profiling that produces legal or similarly significant effects on individuals within the meaning of GDPR Art. 22. All profiling outputs are enrichment signals for human-reviewed or human-configured business logic.

8.4 Data Storage Modes