In a recent LinkedIn post, Yonathan Cohen shares a novel approach to sales analysis, detailing how he leveraged AI to uncover the true reasons behind lost deals. Rather than relying on standard CRM logging, Cohen utilized an AI model to process raw sales data, including call recordings, notes, and emails, to identify patterns and root causes that were previously obscured.
Cohen highlights the common tendency to oversimplify reasons for lost sales, stating, “we mark ‘price’ on everything. The calls said otherwise.” This observation underscores a critical gap between perceived reasons for failure and the actual customer feedback captured during the sales process. By feeding comprehensive data into an AI, Cohen aimed to cut through the noise and identify actionable insights.
The Power of AI in Deep Sales Analysis
The core of Cohen’s strategy involved a specific prompt given to an AI model named Claude. The prompt instructed the AI to analyze six months of lost deal data, including notes, emails, and call recordings, to identify the real reasons for loss, group them, and prioritize fixes. This detailed approach moved beyond superficial data points to a more profound understanding of sales dynamics.
One of the key revelations from the AI analysis, as detailed by Cohen, was the identification of recurring objections. He notes that buyers were consistently using the same phrases to reject deals, a pattern that had gone unnoticed by his team. “nobody had put those calls side by side. Nobody could,” Cohen wrote, emphasizing the scale of data that was previously unmanageable for manual review.
Identifying the ‘Real’ Moment a Deal Dies
Beyond just identifying reasons, Cohen’s AI analysis focused on pinpointing the exact moment a deal lost momentum. This is a crucial distinction from simply noting the date a deal was officially marked as lost. As Cohen explains, the AI sought to identify “The moment the deal died – not the day we marked it lost. The call where it stopped moving.” This granular level of detail allows sales teams to understand critical inflection points and intervene more effectively.
The AI’s output was not just a report but a prioritized list of actions. Cohen emphasized that the AI provided “What to fix first – ranked. Not a report. A decision.” This actionable intelligence is designed to guide immediate improvements rather than offering abstract analysis. Cohen suggests that most teams resort to guesswork when analyzing lost deals, but his approach sought empirical evidence directly from customer interactions.
Attribution to AI-Native Platforms
Cohen attributes the success of this exercise to the capabilities of AI-native platforms, specifically mentioning Attio. He explains that such platforms are built to ingest and process diverse data sources like calls, emails, and notes directly, rather than relying on limited, manually entered CRM fields. “Your CRM already knows why you’re losing. Nobody ever had the time to read it,” Cohen posits, suggesting that AI can bridge the gap between available data and actionable insights.
This method, as presented by Yonathan Cohen, offers a compelling case for integrating advanced AI into sales operations for more accurate and effective deal analysis and improvement.
📝 About This Content
This article is based on insights shared by Yonathan Cohen on LinkedIn.
📅 Originally posted on July 21, 2026 | View original post on LinkedIn →