Cutting Search Time, Streamlining Ops, and Scaling Expertise with GenAI by deepsense.ai x OpenAI for Fennemore 

In a Nutshell

Client’s Challenge

Fennemore, a fully remote and fast-growing Am Law 200 firm, faced bottlenecks due to fragmented knowledge across SQL and SharePoint, slow attorney matching, and inefficient onboarding. Productivity declined, highlighting the need for a unified, intelligent system to integrate unstructured content.

Our Solution

We developed a hybrid GenAI solution powered by ChatGPT Enterprise and the OpenAI API, integrating data from SQL with unstructured content in SharePoint. The system features custom GPT capabilities, intelligent search with RAG, and backend microservices for document retrieval, attorney lookup, and time analysis. The platform was developed in close partnership with the OpenAI Technical Success Team.

Client’s Benefits

The AI solution reduced search time for internal knowledge, improved attorney-to-case matching based on availability and expertise, and accelerated onboarding for new hires.The scalable architecture lays the groundwork for broader GenAI adoption across the firm.

A Deep Dive

Client

Founded in 1885, Fennemore is a full-service business law firm that’s forward-thinking and has a strong track record of growth, including leading the Am Law 200 in revenue increase in 2023. Because of its fully remote model and continued expansion, Fennemore invests in technology to help efficiently scale the firm’s operations.

Overview 

Fragmented Knowledge in a Growing, Remote-First Law Firm

Fennemore faced operational challenges common to law firms, including information fragmentation due to having multiple versions of policies and other crucial documents scattered across various SharePoint drives, as well as the need to efficiently connect cases with attorneys with relevant expertise and time availability. Operating in a fully remote environment and expanding through acquisitions amplified the effect of these complexities. 

Fennemore internally created a proof of concept (PoC) to explore which solutions could enhance operational efficiency, but they needed a partner to build a full-scale system. That’s why they brought in deepsense.ai, a Services Partner of OpenAI.

Faster Access, Smarter Staffing, Stronger Compliance

To address this, the teams at deepsense.ai and OpenAI developed a tailored solution that leverages ChatGPT Enterprise. The goal was to provide Fennemore attorneys and staff with streamlined access to the firm’s collective knowledge, integrating key internal data sources into a smart and easy to use conversational interface. This initiative aimed to streamline daily operations and support the firm’s strategic growth. 

Why deepsense.ai? Proven RAG Excellence in Regulated Environments

We have a proven track record of building LLM-based systems that focus on data privacy, retrieval accuracy, and domain expertise. Our experience in sensitive sectors like legal, healthcare, and finance, combined with proprietary LLM frameworks and observability tools, made us uniquely suited to deliver a secure, scalable RAG solution. The result: a system aligned with a modern law firm’s high compliance and precision standards.

Challenge 

Fennemore’s primary challenges stemmed from: 

• Information Silos: Data critical for daily operations was spread across disparate systems (SQL databases, SharePoint document libraries, website data). 
• Information Access Difficulties: The workforce struggled to find information and identify colleagues with the right expertise.
• Complex Attorney Matching: Identifying the right attorney based on specific expertise, availability, and jurisdiction across a pool of ~350 specialists was a manual, time-consuming process due to scattered knowledge and lack of centralization.
• Onboarding Efficiency: Rapid expansion requires faster ways to integrate new team members and make internal policies readily accessible. 
• Technical Integration: Combining structured data (SQL) with unstructured content (documents) within a unified, intelligent system posed technical hurdles. 

Process

Our approach was deeply collaborative, designed to keep stakeholders engaged and aligned from start to finish.

1. We began with a deep-dive scoping phase, holding daily sessions with the client to prioritize high-impact use cases.
2. Development followed an agile, iterative model, with weekly demos ensuring transparency and ongoing alignment.
3. Educational sessions were conducted throughout to set realistic expectations and build trust in AI capabilities.
4. A structured testing and feedback phase allowed us to fine-tune the solution based on real user input.

This end-to-end engagement ensured the final system was not only technically sound, but also strategically aligned and ready for seamless adoption across the firm.

Solution Overview 

After Fennemore developed a simple internal proof-of-concept (PoC) to test the potential of AI-powered knowledge access, the results confirmed the opportunity for greater operational efficiency. However, turning this early concept into a robust, enterprise-ready solution required deeper expertise, leading them to partner with deepsense.ai, a Services Partner of OpenAI.

Together, deepsense.ai and OpenAI designed an intelligent solution built on ChatGPT Enterprise. The initial plan was to develop a custom chat application using the OpenAI API directly. But since Fennemore already had a ChatGPT subscription, the team pivoted to building a custom GPT within the existing framework, accelerating development while maximizing their current investment.

The deepsense.ai solution features: 

• Hybrid Architecture
  Utilizes Fennemore’s existing ChatGPT Enterprise subscription for the user-facing chat interface (via a custom GPT) combined with a bespoke backend built using the OpenAI API. This maximizes their investment while allowing for tailored data handling. 
• Integrated Data Ecosystem
  Connects to key internal data sources, including SQL databases and SharePoint document libraries.
• Intelligent Search & Retrieval
  Leverages Large Language Models (LLMs) for understanding user queries and retrieving relevant information, whether it’s structured data about attorney availability or content from policy documents. 
• Targeted Functionality
  Designed to specifically address core needs like finding attorneys based on complex criteria, querying time data, and accessing internal guidance documents. 
• Partnership
  Developed in consultation with the OpenAI Technical Success Team to assess and refine the technical approach utilizing GPT capabilities. 

The system employs function calling to direct user requests to appropriate backend microservices designed for tasks like employee lookup, time data analysis, and document retrieval using Retrieval Augmented Generation (RAG). 

Solution Goals 

The solution designed by deepsense.ai was developed with four key goals in mind: 

• Unify Information: Consolidate internal data (e.g., time reporting, people databases, internal policies) into a single cohesive knowledge layer. 
• Improve Accessibility: Enable intuitive, AI-powered access to critical information for a fully remote workforce. 
• Streamline Workflows: Reduce time spent on administrative tasks such as finding attorney expertise, availability, and jurisdiction. 
• Support Expansion: Provide scalable, efficient onboarding and knowledge-sharing tools to support rapid growth and acquisitions. 

Benefits 

This AI-powered solution is designed to: 

• Provide faster and more intuitive access to internal information and expertise.
• Streamline the process of matching attorneys to case needs based on skills, experience, and availability. 
• Simplify access to internal policies and procedures, potentially reducing onboarding time. 
• Offer a scalable platform that supports Fennemore’s ongoing regional expansion and acquisitions. 
• Enhance overall operational efficiency by reducing time spent on administrative search tasks.

Impact & What’s next 

Fennemore has successfully leveraged deepsense.ai’s expertise and OpenAI’s technology to enable seamless information access and significantly reduce workflow challenges. By creating a custom GPT solution, integrated with backend services using the OpenAI API, the firm aims to unify its internal knowledge, making it easily accessible to its remote workforce. 

This initiative is positioned to enhance crucial processes like attorney-case matching and onboarding, providing a scalable foundation to support Fennemore’s continued growth and commitment to innovation. We expect further results and a more detailed impact assessment as the solution scales across the organization.