PulseAgent

LLM Agent Challenge · Tasks A + B

Know your user.
Before they speak.

PulseAgent models how specific users think, write, and choose — simulating authentic reviews and surfacing recommendations that match intent, not just history.

Input — User + Item

User Persona

User ID

Tone Profile

Load Real User from Dataset

Review History

Category

Rating

★ ★ ★ ★ ★

Review Text

Item Details

Item Name

Category

Cuisine / Type

Price Range

🧠

Fill in the form and hit
Simulate Review to see results

Running PersonaConstructionAgent... Building behavioural + textual + contextual profiles ReviewSimAgent inferring rating + generating review

Input — User + Intent

User Persona

User ID

Tone Profile

Load Real User from Dataset

Review History

Category

Rating

★ ★ ★ ★ ★

Review Text

What are you looking for?

🎯

Describe what you're looking for
and hit Get Recommendations

Running ReasoningAgent... Extracting intent · detecting cold-start RankingAgent scoring + injecting diversity