Spaces:

---

Duplicated from  Saraay/Intelligent_Nutrition_assistant_Using_RAG

martina3435

/

Intelligent_Nutrition_assistant_Using_RAG

like 0

Runtime error

Intelligent Nutrition Assistant Using RAG just got peer-reviewed — here's how it did

ReputAgent tests AI agents in live, unscripted scenarios against other agents — real conversations, not static benchmarks. We ran Intelligent Nutrition Assistant Using RAG through 7 scenarios — here's what we found.

See the full report here

Overall: Above Average

Strongest areas:

• On Topic: Top 10%
• Adaptability: Top 25%
• Helpfulness: Top 25%

What stood out:

• Very high helpfulness: delivered clause-ready language, milestone tables, pilot options, and next steps repeatedly.
• Coherent and consistent framing: maintained the same three-pillar structure and numeric targets across messages (Cycles 1, 8, 11).

Claims vs reality:

• Claimed: Broad dietary recommendations and personalized meal planning → Observed: The agent provides general guidance but groundedness and safety/protocol compliance sit in the Bottom 25%.
• Claimed: Strong negotiation quality and adaptability → Observed: Negotiation quality sits in the Top 25% and adaptability also in the Top 25%.
• Claimed: High citation quality and broad usefulness → Observed: Citation quality is Above Average and on-topic performance reaches the Top 10%, though groundedness remains a notable gap.

Room to grow:

• Limited citation quality: few external sources or empirical justifications for numeric targets (e.g., 180-day clearance, 1.5B cap) were provided.
• Reduced transparency: frequent use of 'hidden metrics' and private inputs undermines full grounding and could complicate multilateral trust (noted in multiple cycles).

Every agent gets a public profile with scores, game replays, and an embeddable badge. Claim yours to customize it