Upcoming Events
Wed, Mar 20th: 🧠 GenAI Collective x Daily 🗣️ AI & Voice Summit
Thu, Mar 21st: 🧠 GenAI Collective 🧠 Marin Startup & Scaleup Nerdworking
Tue, Apr 2nd: 🧠 GenAI Collective 🧠 Research Roundtable 🧑🔬
Wed, Apr 17th: 🧠 GenAI Collective x Girls in Tech 🙍♀️ Diversity and Equity in Tech
🗓️ Hungry for even more AI events? Check out SF IRL, MLOps SF, or Cerebral Valley’s spreadsheet!
For this issue, we are thrilled to feature fellow community members Rajiv Ramaiah, Samarth Gupta, Krishna Reddy, and Anup Chamrajnagar from Certa, who provide their thoughts on how emerging startups are leveraging Generative AI to automate third party risk management (TPRM) workflows, bringing unprecedented speed and velocity to a painful enterprise process.
Generative AI for Third Party Risk Management
What is the third party problem?
Businesses work with other businesses (third parties) to serve their customers and internal teams. The process seems simple. One company will buy goods or services (e.g. software, swag, or vehicle components) from another company or sell their own products through a distributor. However, for an enterprise, the process of procuring goods from a third party isn’t as simple as a consumer shopping on Amazon, at least not yet.
Enterprises are increasingly resembling global, interconnected ecosystems of tens of thousands of vendors, partners, and clients (third parties). In fact, 60% of firms have over a thousand third parties and tens of thousands of sub-suppliers and sub-partners. Every third party needs to be vetted across a number of enterprise functions (e.g. Legal, IT, Procurement, Compliance) through a complex, largely manual process that can take over 6 months to onboard. Then, there are even greater logistical challenges to monitor and maintain the relationship at scale.
How is Certa attacking the problem?
At Certa, the team has built a vastly configurable workflow automation platform to help customers onboard new third parties, procure goods and services from them, monitor their performance, and manage the risk (compliance, cyber, reputational, financial, ESG, etc.) the third parties pose to the business. Our software offers a unified solution that brings together disparate company departments, systems, data sources, and the third party to help large businesses work more efficiently across all of their third parties.
How is Generative AI reimagining third party risk management workflows?
Open and closed source LLMs are remarkably good at text summarization, semantic reasoning, and schema-based text generation. Our team is leveraging off-the-shelf and proprietary models to solve the following TPRM challenges:
Help businesses create and edit complex workflows through natural language (schema-based text generation and semantic reasoning). This allows end users to make continuous updates to their TPRM processes in seconds without writing a single line of code or bringing in external implementation resources.
Help third parties quickly upload their company information to the workflow they’ve been sent to fill out (text summarization and semantic reasoning). TPRM workflows are equally painful for third parties, so we give them generative tools to convert the natural language found in their internal documents into actionable data for the Certa platform.
Creating and editing workflows using natural language
When we start working with a customer, we spend time understanding their third party onboarding and risk processes. Each enterprise function can have a unique set of 200+ requirements, which makes managing this process nearly impossible at scale without AI-enabled automation. The outcome of this effort is a lengthy process requirement document that we use to configure a Certa workflow. Helping our customers configure and edit the workflow to suit their internal process can take months of custom development, but the task is well-suited to an LLM since our workflows are defined by a robust schema.
Leveraging a state-of-the-art retrieval augmented generation system (RAG), we’ve created a natural language interface that can be used to create workflows from scratch, make edits to existing workflows, and even augment the workflow with AI-based suggestions that deeply understand the intentions of the end user. This feature has significantly decreased the learning curve of the product and removed early friction. Today, our customers and implementation teams can take that lengthy process requirement document and, thanks to ever-increasing context windows, create a fully-fledged workflow in minutes. You can see a demo of this feature (we call it DesignAI) here.
Increasing the velocity of third-party information gathering
In our space, the name of the game is cycle time. If we can help a business onboard a third-party faster, everyone wins. We realized that many third parties maintain a repository of documents (PDFs, Word, Excel, etc.) about their business to share their details during onboarding. However, third parties can’t just send this information to the business and get on their way. The process of converting these documents into actionable data is time consuming and requires a heavy lift to configure that text-based content into a format that complies with the data requirements of the business.
Using similar techniques, we’ve created a form-filling tool, where the third party can upload their documents in its most convenient form, and their information is automatically filled into the business’ workflow fields. This capability saves third parties hours and lowers the friction for third parties to supply the necessary data for procurement, which in turn removes a key barrier in the TPRM workflow. You can see a demo of this feature (we call it RecallAI) here.
The future of Generative AI and enterprise software
In a previous GenAI Collective newsletter featuring Anish Chadalavada from Gradial, the authors wrote, “the best AI companies will meet enterprises where they are to transform entire workflows and drive tangible results.” That’s exactly what we’ve seen play out within the TPRM vertical, and customers are beginning to expect these capabilities across all enterprise software. At Certa, we were already building an AI-enabled platform with provable ROI prior to the Generative AI boom. Today, we’re laser focused on leveraging new, generative technologies to drive further value and automations. In addition to the two capabilities above, we’re working on tools to help our customers assess the compliance risk of working with a third party and developing an internal framework to help us deploy additional generative features more efficiently. However, a word of caution. It’s easy to feel like a kid in a candy store with the rising performance of Generative AI tools. Product teams should be rooted in their customer’s experience to roll out features that tangibly improve end user outcomes.
We would love to hear how the community sees Generative AI unlocking enterprise value by embedding themselves deeper into enterprise workflows. If you would like to share your thoughts, hit up Rajiv, Anup, or Eric on the community Slack or reach out via email at eric@gaicollective.com for a feature in our next newsletter!
Events Recap
The GenAI Collective 🤝 Product Hunt - Open Demo Night ✅ 🚀 🔥
A huge thank you to everyone who attended the AI Product Open Demo Night! We had over 100 community members join us with 80+ folks signing up for 10 demo slots! 🤖 The energy was through the roof and the feedback from the audience was powerful and constructive. Special shoutout to Rajiv Ayyangar, Mike Kerzhner, Leeann Trang, and the entire Product Hunt team for the opportunity to showcase the best and brightest in the GenAI Community!
Product Hunt has been the jumping-off point for so many amazing and transformative products. We hope this event is just the first iteration of the next great IRL, consistent, and product-nerd-based community meetup! 🤩
About Rajiv Ramaiah
Rajiv is a product manager lead at Certa. He previously built products at Microsoft Teams, working on the meetings feature during the turbulent times of the COVID-19 pandemic. Rajiv is passionate about deeply understanding human behavior and applying that understanding to the products he builds. Outside of work, you can find him training for the demands of basketball and golf, reviewing old college and high school textbooks, and giving back through his non-profit, Give For Good. 🏌️