top of page
Cover.jpg
Frame.png

Multi-turn UI & HITL for Highspot Agents

Developed the fundamental multi-turn user interface, conversational guidelines, and response framework for Highspot Agents. Implemented a reusable Human-in-the-Loop design pattern within Highspot Agent's system workflows, facilitating completion and asset creation through Agents.

ROLE

Agents Platform Crew -
Product Designer

DURATION

Mar '25 - Present

TEAM

Design System - Senior Product Designer (1), Product Manager (1), Agents Platform Crew Engineers (7), AI & ML Crew Engineers (4)

I as a Product Designer for the Agent's Platform crew at Highspot was tasked with creating and optimizing workflows for managing content, facilitating personalized learning, and enhancing guided selling via the Agentic Platform.

HIGHSPOT AGENTS

Highspot Agent is an AI-powered assistant embedded within Highspot that helps go-to-market teams answer questions, take action, and improve performance using real business data.

Highspot's Agent answers questions using trusted data and insights at the right level of granularity, depth, and context. It not just retrieves information, but understands intent and delivers relevant, actionable responses tied to content, training, coaching, and deal activity.

The Agent is expected supports nearly every persona across the go-to-market organization. From reps and frontline managers to enablement and operations leaders, it adapts responses based on the user’s role, goals, and workflow context.

So they can

  • Execute with data and insights

  • Answer questions about the status of content, training, and other assets, and take immediate action inside Highspot.

  • Optimize with data and insights

  • Identify what can be improved across GTM initiatives, uncover performance gaps, and act on recommendations to drive better outcomes.

CHALLENGE

We lacked a unified UX language for how these agents should "act" and "speak". I needed to design the foundation for multi-turn conversations and a robust Human-in-the-Loop (HITL) framework to ensure users felt in control of AI-driven actions.

​

We were designing:

  • A multi agent system inside an enterprise platform

  • A conversational layer that needed to scale into actions

  • A framework that other designers could reuse

The complete process and solution artifacts are not displayed publicly due to confidentiality. Feel free to contact me to view the whole case study

bottom of page