Until recently, use of AI/ML in sales was limited to providing insights and predictions on top of coarse and well-structured CRM data. The advent of Sales Engagement Platforms (SEPs), which orchestrate sales activities across different sales tools and deal stages and servers as the single tool to work from, made it possible to collect much more fine-grained sales activity data, including natural language content of sales conversations across emails, calls, meetings, etc. This new data, coupled with the ability of SEPs to automate activities, extract information, and provide recommendations to sales reps in real time, gave rise to many new and exciting applications of AI/ML in sales.
This talk will present an overview of AI/ML applications in the sales engagement domain. I will discuss both the opportunities and the inherent challenges of applying AI/ML to this area. I will then deep dive into a real-world example of how a leading SEP Outreach.io, used AI/ML to understand the intent behind prospect’s replies to sales rep’s emails, such as willing to meet, more information needed, or not interested, and then suggest next best action.
The problem of intent classification, while well-studied in other domains, is particularly challenging in the sales engagement domain due to very limited training data constrained by security and privacy policies, changing nature of communication at different stages of the deal, and diversity of sales process across different industries and companies. I will discuss how Outreach solved these problems using high performance transfer learning (HPTL) approach to reach the next level of intelligent sales engagement.