When Machine Learning Meets Sales Psychology
Listen to this article:
Conventional sales wisdom treats persistence as virtue: stay in the conversation, overcome objections, keep the line alive. But recent research into the dynamics of sales conversations suggests that our bias toward persistence leads to a massive misallocation of resources. In “Learning When to Quit in Sales Conversations,” a team including Eva Ascarza, Professor of Business Administration at HBS and co-founder of the Customer Intelligence Lab at D^3, explain how they built a generative AI “stopping agent” that watches sales transcripts in real time and chooses to quit or wait to maximize cumulative payoff. The result? The ability to lift expected sales by over 30%.
Key Insight: Quitting is an Optimization Problem
“[W]hile our stopping agents favor quitting early, salespeople hesitate and delay.” [1]
The authors argue that “dynamic qualification”—”the decision of whether and when to quit a sales call that is unlikely to succeed” [2]—is an under-studied decision problem with enormous operational stakes. In a dataset of 11,627 sales calls analyzed by the researchers, over 94% fail, with an average call length of 169 seconds. Humans tend not to be responsive enough to subtle, early linguistic cues that indicate a lack of progress, while simultaneously becoming too responsive to negative signals only after they have already invested significant time in the call. The authors show that early call language is informative enough that a fine-tuned LLM can predict eventual failure quite well after just 60 seconds. The critical challenge is not only in predicting failure, but determining the optimal moment to intervene, which the authors call an ‘optimal stopping’ problem.
Key Insight: Teaching Machines the Wisdom of Hindsight
“Our stopping agent […] identifies subtle linguistic indicators of a lack of conversational progress to quit earlier than salespeople. Further, and unlike salespeople, our stopping agent exhibits dynamic variation in its quitting strategy.” [3]
The researchers’ innovation lies in recognizing that you can retrospectively identify the optimal moment to quit any historical conversation by comparing what actually happened with what would have happened at different quitting times. For each call in the training data, the algorithm calculates when quitting would have maximized the expected cumulative reward (balancing time costs against sales revenue). This creates a dataset of “expert” decisions—the optimal quit-or-wait choice at every moment in every conversation. The researchers then fine-tuned GPT-4.1 to generate these optimal decisions given the evolving transcript. What makes this particularly powerful is how the stopping agent learns dynamic decision rules that evolve as the conversations progress. At 30 seconds, it focuses on whether the salesperson reached the correct person. By 60 seconds, it shifts to gauging interest levels. At 90 seconds, it keys into whether the prospect already has a similar product or service.
Key Insight: Reclaiming Revenue Hidden in “Dead Air”
“These results show that, even under privacy and computational constraints, firms can effectively deploy our stopping agent to improve sales efficiency.” [4]
The managerial implications of this research are profound, moving beyond mere cost-cutting to actual revenue generation. By implementing the “stopping agent” at different levels of “aggressiveness,” the study demonstrated that a firm could reduce the time spent on failed calls by 54%. [5] Crucially, this isn’t just about ending calls faster, it’s about what happens next. When the time saved by quitting doomed calls was reallocated, expected sales increased by up to 37%. This represents a massive gain in productivity. In sales, the focus has often been on “effort motivation”—using commissions and quotas to make people work harder, but this research argues that “effort optimization”—helping people work smarter by knowing when to stop—might be the more powerful lever. The stopping agent also doesn’t replace the human element of persuasion. Instead, it serves as a silent assistant that protects the salesperson’s time.
Why This Matters
For today’s business leaders, the takeaway is clear: efficiency is not just about doing things faster, it’s about choosing not to do the things that don’t work. In an era where AI is increasingly viewed through the lens of total automation, this research also offers a more sophisticated model. It demonstrates that the most effective use of generative AI isn’t to replace the human salesperson, but to provide them with “decision support” that corrects for natural psychological biases. This methodology scales beyond sales to any domain with sequential decisions and observable outcomes. The question for leaders isn’t whether their teams face similar cognitive constraints, they almost certainly do, it’s whether they’re ready to systematically identify and correct them.
Bonus
For another use case where AI doesn’t replace humans, but offers the opportunity to improve judgment, break silos, and accelerate execution, check out “The Cybernetic Teammate: How AI is Reshaping Collaboration and Expertise in the Workplace.”
References
[1] Manzoor, Emaad, Eva Ascarza, and Oded Netzer, “Learning When to Quit in Sales Conversations,” arXiv preprint arXiv:2511.01181 (2025): 23. https://doi.org/10.48550/arXiv.2511.01181. See also https://stoppingagents.com/.
[2] Manzoor et al., “Learning When to Quit in Sales Conversations,” 1.
[3] Manzoor et al., “Learning When to Quit in Sales Conversations,” 2.
[4] Manzoor et al., “Learning When to Quit in Sales Conversations,” 21.
[5] Manzoor et al., “Learning When to Quit in Sales Conversations,” 2.
Meet the Authors
Emaad Manzoor is an Assistant Professor of Marketing and a graduate field member of Computer Science at Cornell University.
Eva Ascarza is Professor of Business Administration at Harvard Business School and co-founder of the Customer Intelligence Lab at D^3.
Oded Netzer is the Vice Dean of Research and the Arthur J. Samberg Professor of Business at Columbia Business School.