of engaged service providers each week originate from the recommendation engine's suggestions
delivered within Salesforce for every new concierge request, enabling fast provider selection
recommendation engine now used for all incoming concierge requests across the platform
Our client connects private equity (PE) firms with service providers (SPs) for tasks such as diligence, value creation, and prep-for-sale, offers a "white-glove" concierge service. Their challenge was speeding up the process of selecting the best SPs from their large network to match new client needs, ensuring high-quality, timely recommendations for every new concierge.
OneSix built a recommendation engine using a domain-general large language model (LLM) to extract, distill, and embed unstructured data about past concierges and SP qualifications into a searchable vector database. This system allows agents to quickly search for and find the most relevant past concierges and SPs.
To enhance the relevance of the search results, a proprietary re-ranking model was included, ensuring that the recommendations are hyper-relevant and aligned with the client’s core sales metric. The engine was integrated into their Salesforce frontend via an API, and also exposed a natural language “chatbot” interface for real-time recommendations based on user input data.
Additionally, the system leverages existing data from Salesforce, Fivetran, and Snowflake, creating a specialized data-science schema to feed the recommendation engine. The Python app, connected to Snowflake via the Snowflake Python connector, powers the entire solution.
The recommendation engine is now used for all incoming concierge requests. Within Salesforce, agents receive a top-10 list of recommended SPs for each new concierge, allowing them to quickly choose the best-fit providers.
Success is measured by the percentage of engaged SPs that originated from the system’s recommendations, typically ranging between 60-80% each week. The solution has significantly streamlined the company’s SP selection process, delivering fast, high-quality matches to clients.
More to explore
OneSix created a Power BI solution for a global restaurant chain, enabling interactive, on-demand insights into sales performance and customer sentiment across international markets.
OneSix developed a personalized, data-driven marketing platform for a real estate client, driving increased revenue, customer engagement, and establishing a scalable foundation for long-term, automated personalization.
OneSix implemented machine learning models for a financial services client to optimize marketing spend, resulting in a strategic reallocation that improved ROAS from 0.5-0.8x to 1.5x and achieved the client’s first quarter of positive marketing ROI.