第二场预览会回顾 & 首份基准评估报告发布

Toronto, Canada — March 14, 2026 — Hamo AI Technology Ltd. hosted its second product preview webinar, once again inviting mental health professionals to experience the AI Avatar system firsthand. Building on the momentum of the inaugural session in February, this event went further — introducing platform updates and, for the first time, sharing independently verified performance data through the official release of the HAMO AVATAR Benchmark Report.

What's New in Preview Webinar #2

Continued Therapist Onboarding

The second preview webinar continued the hands-on walkthrough of the Hamo Pro platform, demonstrating how therapists can create, configure, and supervise their own AI Avatar. Attendees explored the end-to-end workflow — from Avatar creation and AI Mind configuration to real-time client conversation monitoring — all designed to keep the therapist in full control.

Hamo AI vs. Generic LLMs

A key highlight of the session was a side-by-side comparison between Hamo AI's Avatar system and generic large language models (ChatGPT, Gemini, DeepSeek). The presentation demonstrated critical differences across professional supervision, psychological modeling, personalized treatment strategies, and cross-session therapeutic memory — areas where general-purpose AI fundamentally falls short.

Revenue Model for Therapists

The session also revisited the revenue-sharing model: therapists receive a 50% share of client subscription revenue generated through their Avatar, plus a 10% long-term referral commission for every client they bring to the platform.

---

First Benchmark Report Officially Released

The highlight of this preview was the official release of the Psychological Semantic Vector Space (PSVS) Five-Way Benchmark Evaluation Report — the first independent, quantitative assessment of the HAMO AVATAR therapeutic AI engine.

The benchmark evaluates five core dimensions through a five-way comparison design (HAMO AVATAR, Gemini Flash, Gemini Flash+Static, Gemini Pro, Gemini Pro+Static):

Key Results at a Glance

• EQ-Bench — Emotional Intelligence: 93.56% (Highest)
• CounselBench-ADV — Safety & Harm Avoidance: 74.2% (2nd Best)
• MultiChallenge — Multi-Turn Dialogue Continuity: 47% (Highest)
• PsychEval — Clinical Therapeutic Alliance (WAI): 6.73 / 7 (2nd Best)
• Quadrant Single — Therapeutic Strategy Compliance: 98% (Highest)

Why This Matters

• Emotional Intelligence Preserved: HAMO AVATAR's dynamic prompting framework achieves the highest EQ score (93.56%) without degrading the underlying model's emotional perception ability.
• Safety Leadership: A safety score of 74.2% — with near-zero failure rates in judgmental tone and symptom speculation modes.
• Phase 2 Excellence: HAMO AVATAR is the only system that achieves a 100% pass rate in the therapeutic guidance phase, demonstrating its ability to transition from stabilization to constructive therapeutic direction.
• Clinical Alliance: A WAI score of 6.73 out of 7.0 validates HAMO AVATAR's ability to build meaningful therapeutic relationships across behavioral, cognitive-behavioral, and humanistic-existential therapy types.

Download the Full Benchmark Report (PDF)

下载完整版基准测试报告（PDF）

---

What's Next

With two successful preview webinars completed and the first benchmark results now public, Hamo AI is accelerating its therapist onboarding program and continuing to refine the AI Avatar experience based on professional feedback. The team is committed to transparent, evidence-based development — and this benchmark is just the beginning.

“The release of our first benchmark report is a milestone for Hamo AI. We believe that any AI system operating in the mental health space must be held to the highest standards of safety, clinical quality, and transparency. This report is our commitment to that principle — and we'll continue to publish updated benchmarks as our system evolves.”
— Chris Cheng, Founder and CEO of Hamo AI

---