Fast Insights: Qualitative Analysis of Scam Networks

Evidano is an AI-powered qualitative data analysis platform that ingests mixed-format evidence, automates thematic coding, and links textual, infrastructure, and financial signals.

AP’s June 30, 2026 investigation found U.S.-based hosting, AI models (chiefly ChatGPT and Gemini), and Starlink connectivity were used to scale scams, and analysts can reproduce insights by coding transcripts, device logs, and blockchain traces.

• AP sampled 202, 013 device connections (Feb 2025–Jan 2026) and reported approximately 50, 000 victims in one month.
• Estimated U.S. losses cited by AP include nearly $200 billion (FTC estimate) and at least $75 million in illicit network-level profits (TRM Labs / AP reporting).
• Use a seven-step checklist to collect transcripts, device/IP logs, and blockchain IDs, run transcription and auto-coding, validate codes, and produce visual evidence for policy or takedown actions.

Fast take: AP’s June 30, 2026 investigation found AI models, U.S.-based hosting, and Starlink connectivity were used to scale scams from Myanmar to dozens of countries.

Read the original reporting here: AP / FRONTLINE investigation.

• Why analysts should care: the report links textual artifacts (scripts, chat logs), infrastructure signals (202, 013 device connections analyzed) and financial traces (crypto flows), exactly the heterogeneous corpus qualitative research tools must ingest and code.
• Payoff: we show a pragmatic workflow to extract themes, compare segments, and produce visualizations you can share with policymakers or platform safety teams.

What happened: AP and FRONTLINE combined leaked scam-center files, satellite imagery, 58 victim interviews, device connection logs, C4ADS analysis, and TRM Labs blockchain work to show a multi-layered abuse chain.

• Tools used: ChatGPT and Google Gemini integrated into scam suites to produce automated replies, role-play bots, and more than 100-language translation.
• Infrastructure role: one in five device signals from sampled scam compounds routed via U.S.-registered providers, and Starlink was the top ISP in Myanmar during sampled periods.
• Evidence types analysts can collect: chat transcripts, platform account logs, device/IP mappings, satellite imagery timestamps, and blockchain transaction traces.

Implications for researchers and analysts: combine thematic coding with infrastructure and financial traces to produce actionable findings for product, policy, or enforcement.

Do more, faster with Evidano: use a single workspace to ingest mixed-format evidence, run automated thematic coding, and map texts to infrastructure and blockchain signals.

• Evidano transcribes audio, applies custom dictionaries (names, slang), and translates content while preserving original timestamps and metadata when you upload leaked files, chat transcripts, interview audio, device logs, and CSVs.
• Run thematic coding across thousands of documents, then compare themes by segment (victim country, scam type, platform) and export frequency tables and codebooks to support reports or legal briefs.
• Map IP/device fields and blockchain identifiers into the same workspace, run cross-variable queries (for example, show transcripts where IP X appears and theme Y is present), and visualize co-occurrence networks for briefings.
• Data is encrypted, stored privately, and not used to train third-party models, and you can use AI chat over your corpus to iterate hypotheses without exposing data externally.

Checklist: follow these seven steps to reproduce the core insights from the AP investigation and extend them for your stakeholders.

• 1) Collect: centralize transcripts, leaked files, device/IP logs, and blockchain tx IDs.
• 2) Clean & ingest: upload to Evidano; run transcription and translation with a custom dictionary.
• 3) Auto-code: apply an initial codebook (scripts, persuasion tactics, infrastructure signals) and let Evidano suggest subcodes.
• 4) Validate: spot-check auto-codes, refine the codebook, and lock for reproducibility.
• 5) Cross-segment queries: compare themes by country, platform, or time window (for example, pre/post Starlink cutoffs).
• 6) Visualize: generate co-occurrence networks and hierarchical theme charts for briefings.
• 7) Produce evidence pack: export coded quotes, timeline visuals, and a brief for legal and policy teams.

Wrapping up: AP’s June 30, 2026 report shows that combining qualitative evidence with infrastructure and financial traces creates the strongest narratives for disruption and policy.

• Try a pilot: ingest one scam-center dataset, run thematic and cross-segment analysis, and produce a stakeholder brief in two weeks. Try Evidano for free.