A leading recruitment agency was struggling to manage its sprawling digital infrastructure, with communication breakdowns and missed opportunities becoming increasingly common. Their core metric—expense-to-placement ratio—was far below target, hovering around 5% instead of the desired 16%. The operational side of the business was stable, but exposure to a volatile job market made consistent performance difficult. The agency needed a centralized system to consolidate digital assets, improve internal communication, and ensure agents could identify and pursue long-tail opportunities more effectively.

The agency set out to better understand and improve its core KPI: the expense-to-placement ratio. Their primary goal was to identify the qualitative factors contributing to underperformance, particularly around how agents utilized digital tools. By focusing on data entry, years of experience, and placement completion, they aimed to assess whether resources were being used effectively to match candidates with roles. The overarching intent was to support their ethos of individual entrepreneurialism—empowering agents with the right tools and insights. However, early data revealed no meaningful correlation between CRM usage and successful placements, prompting a deeper evaluation of the software’s real-world impact.

In response to the lack of correlation between data entry and successful placements, the agency initiated a series of A/B tests to evaluate changes in the CRM’s user interface. Early experiments focused on improving the data entry experience with small adjustments like dropdowns and guided inputs. These incremental changes led to modest improvements, but adoption rates for new features rarely exceeded 80%. Larger updates followed—such as interface hydration and mobile optimization—but these too failed to drive widespread engagement. Recognizing the limits of incremental changes, the agency shifted focus toward understanding the deeper reasons agents found the tool inadequate.

While incremental improvements led to better data entry rates, the CRM still struggled to deliver the operational efficiencies the agency had hoped for. The turning point came with an unexpected insight: agents had begun independently using large language models (LLMs) to strategize, brainstorm, and deepen their understanding of client occupations. This shift revealed that agents were not resistant to digital tools—they simply needed ones that supported their autonomy rather than imposed constraints. As a result, the CRM’s role evolved from being solely a tracking tool to a supportive platform, with increased investment in LLM-powered features aimed at enhancing agent insight and performance.