Sales Engagement vendor Outreach is teasing a new predictive analytics capability called Amplify which leverages the history of a firm’s sequences and workflows. The firm will not be employing a black-box AI strategy but providing recommendations with explanations.
CEO Manny Medina faults deep learning strategies which lack “the ability to make inferences, such as the ability to figure out why things work” and require users to trust the recommendations without providing a basis for the suggestions.
“We believe we need to tackle this problem following general scientific principles. Hypotheses need to be testable, data should be very carefully examined to verify the quality of the data.”
- Yifei Huang, Machine Learning Lead, Outreach
“When we built Amplify, we built it with the core belief in mind that, the human needs to understand why things work so that machine can understand why things work so that the machine can get better at helping the human,” said Medina.
For example, Amplify deploys natural language processing (NLP) around email responses to help identify whether responses are unsubscribes, objections, or positive. Outreach claims that their NLP classification is 92% accurate, only three points behind manual classification.
NLP will also be used to assess objection handling to identify reps who handle objections well and which ones need improvement. This feedback is then available to managers to assist with coaching.
Amplify addresses two key managerial questions: “Is my team adopting the new technology? Is the new technology delivering a measurable lift?”
Amplify will be unveiled at their May Unleash conference.