Kaufman Rossin avoids “AI quicksand” and accelerates mission-critical work with AI Studio

Kaufman Rossin, a financial services firm that provides audit, tax, ‌business, risk, and forensic advisory services, wanted to use generative AI to help employees maximize efficiency and effectiveness in knowledge work, such as research and analysis. They initially considered building their own DIY generative AI solution with vector databases, but ultimately found it too costly and difficult to maintain. 

Kaufman Rossin selected Writer for its graph-based retrieval augmented generation (RAG) technology, security standards, and flexible approach to developing custom AI apps in AI Studio. Sales, security, marketing, and product + UX teams use Writer to accelerate workflows, reduce friction, and tackle more strategic tasks, improving work-life balance.

Tell us about yourself and what you do at Kaufman Rossin.

My name is Albert Primo, and I’m the Chief Innovation Officer at Kaufman Rossin. Kaufman Rossin was founded by a visionary duo in 1962 to be a full-stack advisory firm that offers services such as tax, audit, risk advisory, wealth management, and fund administration. We deliver world-class services to our clients with our expertise, and our people-first culture has and always will power that mission.

As a technical leader, what do you think sets Writer apart from other solutions?

One of the biggest challenges with AI is figuring out how to deal with the hype. We needed to make sure generative AI delivered measurable value‌ — ‌higher productivity and better experiences. To do that, choosing the right generative AI partner is essential.

We initially tried a DIY approach using vector databases to store and retrieve our key company data, but it became very costly and difficult to maintain, and it just felt like we were in AI quicksand. The DIY approach limits functionality because you invest so much time in a specific solution that it becomes difficult to be agile and respond nimbly to any problems that arise.

With the Writer full-stack platform, RAG is already integrated, making it easier to start right away. Writer Knowledge Graph supports different file types and allows us to easily scale and continuously update enterprise data. With Writer, we see depth of functionality, value, and, ultimately, impact.

Writer AI Studio was also a huge differentiator. Its low-code and no-code tools make it incredibly easy to build your own custom AI apps for very specific use cases — you can even connect it to Knowledge Graph to query your own data quickly. AI Studio allows users of all skill levels to leverage generative AI in their workflows. For example, Jason Reed, one of our admin supervisors for the innovation team, helped build most of our AI Studio apps with no engineering background.

What are some of your use cases with Writer?

We started using Writer for research, then expanded to more complex knowledge work, helping employees move up the value chain, develop business, and deliver greater value to clients. AI Studio helps us get to this deeper level of customization and innovation because we can now create AI apps for specific workflows.

In the first three months with Writer, we developed about 10 different AI apps using AI Studio. AI Studio allows employees with the most context to build and test AI apps in a no-code environment. Some of the AI apps we built help us with research, and some help us create new materials for our teams. For example, we’ve built AI apps that retrieve reports for our business valuations team, support our accounting team with FAVS research, create overviews of financial agreements, provide sales pitchbook data, and draft IRS tax notice responses. Some of our most impactful use cases include the GAAP research tool and the due diligence questionnaire app, or DDQ app.

Tell us more about the GAAP research tool and DDQ app.

The Generally Accepted Accounting Principles (GAAP) research tool was the first AI app we developed with Writer. Our auditors faced significant friction with existing tools, going through 20 or 30 clicks to double-check or look up an accounting standard. In some cases, they’d go straight to Google because it was so much faster, but those results aren’t always verified.

With Writer, we set up a Knowledge Graph that powers an app where auditors can get quick, verified answers that show the corresponding sources. We see around 50-70% time savings for key use cases with the new GAAP research tool.

The due diligence questionnaire (DDQ) app was built with the help of Writer to serve our information security team. In the last few years, we’ve been flooded with due diligence questionnaires, and it takes a lot of time to answer each one. Our Chief Information Security Officer has been responsible for almost all of them.

Writer helped us take all the historical data from previously completed due diligence questionnaires and upload them into a Knowledge Graph. Now, when we receive a new request, we can run it through the DDQ app that’s connected to this Knowledge Graph. It generates an output that’s consistent with previous responses and significantly accelerates this process.

What impact have you seen with Writer?

From a ROI perspective, productivity is key. Technology should reduce friction and improve employee experiences, not complicate them.

Writer has also been a catalyst for us to review our internal processes. For each use case, we think of it as a deployment framework. We identify an initial issue, develop an AI app as a solution, and monitor its impact and usage over time to ensure it delivers value and find areas of improvement. Then, we draft recommendations on how we can increase the impact in future iterations.

It’s a process every company should go through if they want to innovate. People are now understanding what AI can do and using it to reevaluate existing workflows, boosting productivity and helping employees focus on higher-value tasks so they can enjoy a better quality of work-life.

How are you driving adoption?

We currently have four teams using Writer, but the goal is to empower all of our employees with Writer. Initially, our Writer AI apps were spread across different locations and hard to find. We wanted our engineering team to build a hub, but that would have taken three to six months and diverted resources.

Instead, we embedded all of our AI apps into a SharePoint site that we built — which we call our “AI app store” — in just two weeks. The apps are now on one landing page, sorted by category, making them easier to find.

From the start, we wanted to be strategic in integrating and scaling generative AI, beginning with engagement and education. Employees can’t use generative AI effectively if they don’t understand it. We held webinars on generative AI and explained potential real-world use cases, then organized an AI use case contest, which spurred engagement and identified over 50 potential use cases.

Additionally, the Writer customer success team has been essential in driving  adoption. They’re so engaged that it feels like they’re a key part of our team and a strategic partner.

What advice do you have for other AI leaders?

For other leaders considering a DIY approach to building their own AI stack, I’d say “think twice.” We’ve gone down that road for about 10 years with our software team taking the “maker” approach, but it quickly becomes very expensive and hard to maintain. Additionally, it’s hard to take another path when you’ve invested so much time and money into a specific DIY approach. Similarly, vertical-specific AI solutions can meet our business needs but will likely lack the flexibility a full-stack solution can offer.

My other advice is that it’s important to adopt the right mindset when it comes to generative AI. Be ambitious but patient. Stay persistent and consistent in your efforts, and be honest with what AI can and can’t do today.