Sally

Sally is an intelligent training agent built to accelerate the learning curve of Key Account Managers (KAM) through advanced simulation, adaptive coaching, and strategic evaluation. The system acts as a digital sales trainer capable of running negotiation scenarios, validating commercial strategies, correcting common errors, and proposing tactical improvements based on B2B best practices.
The working prototype is stable and demonstrates consistent multimodal performance by combining conversational intelligence, structured evaluation logic, and tool orchestration. It leverages voice capabilities, browser automation, and cloud-based agents to deliver an integrated coaching experience. The agent dynamically adjusts difficulty, analyzes user inputs, and generates actionable feedback aimed at improving objection handling, value articulation, closing skills, and account expansion strategies.
The technical stack aligns directly with the hackathon theme by turning voices, workflows, and cloud tools into a cohesive, orchestrated agent ecosystem. It integrates voice synthesis, automated pipelines, sub-agent collaboration, and real-time orchestration to deliver a scalable and high-impact training solution. The project showcases innovation by replicating the reasoning style of a senior KAM, enabling organizations to standardize onboarding, reduce dependence on manual training, and scale sales enablement with minimal friction.

All components were developed during the hackathon. The team worked in parallel on sub-agents, the front-end interface, and the backend architecture required to interconnect every agent. After that, the full system integration with the front-end was completed.