How Prospectiqa Works — Methodology & Standards
Transparency about how we enrich data, how we measure accuracy, and how we write about competitors. Published for users, researchers, and journalists.
Enrichment Methodology
Prospectiqa uses large language model AI (LLM) agents with real-time internet access to research B2B contact data. This is fundamentally different from querying a static database — every enrichment request triggers a live research session. Here is how it works step by step:
1. Multi-source real-time web research
For every contact record, Prospectiqa's AI simultaneously searches across 14+ source categories: company websites, LinkedIn-indexed profiles, professional directories, press releases, speaker bios, author pages, GitHub profiles, conference databases, industry association pages, and public regulatory filings. Unlike static databases that are queried from a pre-built index, every enrichment is a live research session.
2. Email pattern construction & cross-validation
The AI identifies known email addresses at the target company from any found source, infers the company's email format pattern (e.g. firstname.lastname@company.com), and applies the pattern to the target contact. It then cross-validates against any direct references to the contact's email found in public web content. Multiple pattern candidates are ranked by probability.
3. Confidence scoring
Each result is assigned a confidence score from 0 to 1. A score of 0.9+ indicates the email was found directly (verbatim match in public content). A score of 0.7–0.89 indicates a pattern-inferred email cross-validated against at least one corroborating source. A score below 0.7 is marked "partial" and indicates a best-estimate with lower certainty. We never suppress low-confidence results — we surface them with transparent scores so users can make informed decisions.
4. Phone number research
Phone enrichment searches for direct mobile numbers, direct work lines, and company switchboard numbers — in that priority order. Direct numbers are sourced from: public professional profiles, contact forms that include direct lines, speaker bio pages, press release contact sections, and professional directory listings. Switchboard numbers are retrieved as a fallback. Phone numbers undergo format validation and duplicate removal before being returned.
Accuracy Measurement
What Prospectiqa does NOT do
- ✕ We do not store or resell enriched contact data. Each enrichment session is ephemeral.
- ✕ We do not scrape LinkedIn in violation of their terms of service. We research LinkedIn-indexed public web content.
- ✕ We do not guarantee deliverability. Enriched emails are research-based estimates, not verified deliverable addresses.
- ✕ We do not enrich consumer (B2C) data. Our service is designed exclusively for professional B2B contact research.
- ✕ We do not use personal data as training data for AI models.
Editorial Standards
Our content — comparisons, benchmarks, and guides — is written to be genuinely useful, not just to rank. We hold ourselves to the following standards:
Data Sourcing
Prospectiqa's AI researches publicly available information on the web. The types of sources it searches include:
See the methodology in action
Upload your contact list and see how AI real-time enrichment compares to static database tools.
Start Free Enrichment