Company
Highspot's sales enablement platform helps revenue teams improve performance through content management, training, and analytics. By centralizing sales materials and providing data-driven insights, the company enables organizations to increase win rates and accelerate deal velocity.
The Challenge
While 91% of sales managers believe coaching improves performance, only 26% of sales reps receive weekly 1-on-1 coaching (Source: Mindtickle, Highspot). Highspot aimed to bridge this gap by integrating AI-powered coaching directly into their platform, automating feedback on sales pitches that would typically require human review. The technical challenge was significant: their system relied on a legacy Backbone/Ruby architecture with complex custom software that was difficult to navigate, yet critical to the product. The engineering challenge lay in adding sophisticated AI capabilities while maintaining the reliability and performance their customers depended on.
The Solution
A.Team network member Andrew Walsh, a senior frontend engineer, brought the technical expertise needed to integrate AI seamlessly into Highspot's enterprise platform. Walsh enhanced the existing Ruby/Backbone architecture with generative AI capabilities that analyzed both text and video submissions, providing nuanced feedback on pitch quality, tone, and pacing. He leveraged LLMs and Whisper for intelligent analysis while using React components and event queues to ensure the AI features felt native and performed at enterprise scale. Throughout the project, Walsh collaborated closely with Highspot's design team and mentored internal engineers, ensuring knowledge transfer alongside delivery.
Technologies used:
React & TypeScript: Built new AI feedback interface components while maintaining compatibility with existing UI patterns
Backbone: Integrated modern features into the legacy framework without disrupting core functionality
Ruby on Rails: Extended backend APIs to support AI model integration and feedback processing
Python: Algorithm development and custom processing logic for computer vision and anomaly detection
Event Queues: Managed asynchronous processing of video/text submissions and AI response handling
LLMs & Whisper: Analyzed sales pitches for content quality, tone, and pacing to generate coaching insights
"I was deep in the weeds from the start, and soon found myself involved in a lot."
Andrew Walsh, A.Team Frontend Engineer
The Results
Highspot transformed sales coaching from a limited resource into an unlimited capability, fundamentally changing what their platform could offer to clients.
Enhanced Coaching Capacity
Eliminated the need for human reviewers in the coaching loop, allowing clients to provide coaching to all their sales reps.
Cost Efficiency
Reduced operational costs for end users by automating what was previously a manual, labor-intensive process requiring dedicated coaching staff.
Enhanced User Experience
Analytics showed meaningful improvements in user satisfaction, with the AI coaching becoming a valued part of the platform experience
Competitive Advantage
Successfully embedded AI throughout their product, establishing Highspot as an innovation leader in sales enablement.

