HerEthical provides a trauma‑informed AI platform that scans text across documents, emails, chats and case files to detect harmful language such as victim‑blaming, coercive control and fraud cues. The solution offers a web app for on‑demand analysis and enterprise APIs that deliver categorized insights and analytics dashboards, helping organizations improve empathy, reduce regulatory risk, and lower investigation costs.
Funding
Funding not disclosed
Founders
Product
Problem
Organizations that handle sensitive communications—such as courts, police, health services, banks, and charities—often embed subtle victim‑blaming, coercive or abusive language in reports, judgments, case notes and policies. These patterns are difficult to detect at scale, leading to reduced trust, increased trauma for survivors, and higher operational and regulatory risk.
Solution
HerEthical AI provides a trauma‑informed, multi‑agent platform that analyses text across documents, emails, chat logs and case files to surface harmful language patterns such as victim‑blaming, coercive control and fraud cues. Users can upload individual PDFs or integrate the solution via APIs into existing document‑management and case‑work systems, receiving instant, research‑backed analyses that highlight specific phrases and categorize them by type. The platform aggregates results into dashboards, enabling teams to track the prevalence of harmful language by service, team or region and to measure the impact of training, policy changes or supervision models over time. By delivering clear, actionable signals, HerEthical AI helps professionals focus on empathetic response while reducing investigation time, regulatory exposure and financial loss.
Target Audience
Primary customers are public and private sector organizations that produce or review large volumes of textual records—such as courts, law enforcement agencies, health trusts, financial institutions and national charities—who need to ensure empathetic, compliant communication.
Features
- Web‑app for on‑demand analysis of PDFs or pasted text with instant, research‑validated victim‑blaming detection
- Enterprise API for automated scanning of documents, case‑work tools, data lakes and document‑management systems
- Multi‑agent AI that links information across messages, notes and reports to identify subtle coercive or abusive patterns beyond simple keyword matching
- Categorized taxonomy (e.g., minimising, behavioural blame, gaslighting) with examples to support human review and decision‑making
- Aggregated analytics dashboards showing prevalence by service, team, region or document type and before‑after comparisons for culture‑change initiatives
- Privacy‑preserving deployment options, including on‑premise model execution so sensitive text never leaves the organization’s infrastructure
- Customizable category weighting and integration with existing training or supervision frameworks