A national homebuilder was struggling with an inefficient lead generation process. They received a daily list of approximately 2,000 leads with no prioritization or context about customer engagement with their website. This lack of insight meant sales managers had to manually sort through the list, essentially calling people at random, resulting in low conversion rates and wasted time.
The technical challenge stemmed from the data's unstructured nature. Key information about customer interactions was buried within messy JSON arrays and scattered across different systems, including their website and CRM. This data was technically available but was not being properly aggregated or analyzed to provide actionable insights. The client also lacked the in-house skill set to integrate and analyze this complex data to optimize the sales process.
Trility's approach was a two-day workshop that trained the client's marketing and IT teams on Microsoft Copilot. This interactive format allowed us to brainstorm use cases and immediately assess technical feasibility. The most promising idea was to improve lead management, which involved a daily data extraction process that resulted in an output only readable (thus usable) by a Large Language Model (LLM).
Our team collaborated with the client’s team to create a proof of concept. This involved enriching the data extract with meaningful context, including CRM comments and detailed website interaction data. Trility then engineered a custom AI agent within Copilot to process the complex file and demonstrate its potential to:
Apply a weighted score (based on factors like site visits and CRM activity) to identify the top 10 leads.
Create personalized talking points, utilizing a lead’s online behavior and history.
Deliver a clean, usable output, including names, emails, and suggested talking points, for a targeted daily call list.
The workshop successfully demonstrated the tool’s power and provided the client with knowledge, confidence, and a clear path forward. Trility proved that a viable solution was achievable, transforming a previously skeptical chief marketing officer into a project champion.
The primary outcomes were foundational:
Validated an AI agent could process complex data and generate a prioritized, contextualized lead list.
Positive feedback from the sales team and their desire for continued development confirmed the project’s value.
Provided the client's leadership with a tangible example of AI’s potential to drive revenue, empowering them to confidently continue investing in and exploring these solutions.
