Saifr is a regulatory technology company within Fidelity Investment’s innovation incubator that is committed to safeguarding organizations from pervasive AI and regulatory risks. Pulse 2.0 interviewed Saifr co-founder and CEO Vall Herard to learn more about the company.
Vall Herard’s Background
What is Vall Herard’s background? Herard said:
“Currently, I’m the founder and CEO of Saifr, a RegTech within Fidelity Investments’ innovation incubator, Fidelity Labs. We use cutting-edge AI technology to help our clients tackle complex regulatory and compliance challenges such as marketing content review and adverse media screening and monitoring. We benefit our clients by saving time, reducing costs, and improving accuracy while protecting their firms from brand, reputational, and regulatory risks.”
“My career journey has been anything but linear. While I initially studied mechanical engineering and economics, I found my true calling in finance. I found my footing on Wall Street, analyzing investment products for some of the industry’s biggest names. That sparked my interest in risk management, leading me to consult for global firms and implement sophisticated risk systems, especially in the realm of financial market risk.”
“I eventually transitioned into product marketing and sales roles at several successful FinTech companies, often focusing on risk mitigation. These experiences laid the groundwork for my current venture, at the forefront of applying AI to regulatory compliance.”
Formation Of Saifr
How did the idea for Saifr come together? Herard shared:
“The Saifr journey began in 2020, but its roots can be traced back to a pivotal compliance conference I attended earlier. In one of the keynote sessions, I asked the compliance professionals in the audience if they would be interested in using AI to address some of their challenges. The overwhelming, resounding answer was ‘YES!’”
“Yet, they all said it was impossible since the data necessary to train the models didn’t exist. They realized that AI would need a massive amount of data—otherwise, the models’ outputs would be inaccurate. This session confirmed there was a need to bring sophisticated AI to compliance, and I knew where to find the needed data: Fidelity. With that question, the seed of Saifr was planted.”
“Following this experience, I joined Fidelity, built a small team, and developed initial business cases for SaifrReview and SaifrScan, leveraging our unique access to millions of proprietary documents that represented the work of thousands of marketing and compliance experts across various financial sectors. That same year, we created a prototype, gathered user feedback, and with positive experiences from prospects, initiated the engineering work.”
“In January 2022, we officially launched, offering marketing, legal, and compliance teams within financial services a way to streamline their content creation, approval, and filing processes. As we evolved, we enhanced our functionality, enabling our AI to ‘understand’ content as it’s created, highlight potential marketing and compliance risks in public-facing materials (text, images, video, and audio), explain flagged issues, propose alternative language, and suggest relevant disclosures.”
Core Products
What are the company’s core products and features? Herard explained:“Since our launch in 2022, we’ve revolutionized how firms handle content creation, compliance, and risk management. Our flagship products, SaifrReview and SaifrScan, harness the power of AI to analyze marketing content across various media formats, including video, catching 90% of the potential risks that a human would identify. But we don’t just flag possible issues—we explain them, offer alternatives, and suggest appropriate disclosures.”
“Clients can access our models via our workflow tool, SaifrReview, or via SaifrScan (add-ins and APIs that seamlessly integrate into existing systems) to achieve actionable results. They’re seeing substantial cost savings, getting to market ten times faster, and reducing risk exposure.”
“I’m particularly excited about our latest offering, SaifrScreen, which we unveiled in May of this year. This AML/KYC compliance offering takes a novel approach to using AI for unstructured data analysis. Unlike traditional adverse media screening tools that rely solely on structured data such as sanctions, wanted, watch lists, etc. (only 20% of global internet data), SaifrScreen digs deeper by also including unstructured data like news, social media, and government records, delivering a more comprehensive risk assessment.”
“Our layered AI models work in concert to understand the nuances and sentiment in these diverse sources, employing advanced entity recognition techniques to accurately identify and disambiguate specific individuals or organizations. Our comprehensive approach, using both structured and unstructured data, results in almost real-time results, more accurate risk identification, fewer false positives, and a higher degree of confidence that we’ve correctly identified the person or entity of interest.”
Challenges Faced
What challenges have Herard and the team faced in building the company? Herard acknowledged:
“The rapid advancement of AI presents both opportunities and challenges in the financial services industry. Because AI is relatively new, opinions vary widely—some view it as a perfect solution, while others see it as potentially dangerous. At Saifr, our challenge is to strike the right balance: harnessing AI’s potential while addressing valid concerns and always including humans in the loop.”
“At Saifr, we’re working towards more and more complex AI capabilities to help solve compliance challenges, but we know that human involvement remains critical. Even the best AI models can occasionally make mistakes, so we focus on using AI to augment rather than replace human decision-making.”
“We also involve humans in every stage of development and implementation. Our data, which is both proprietary and augmented from other independent sources, has been validated by former regulatory staff attorneys. As we develop models, the outputs are further validated by subject matter experts. Our tools also incorporate the ability for users to provide feedback that we use to continuously improve our models.”
“One of Saifr’s fundamental tenets is that we appreciate AI’s potential while staying grounded in its current limitations. This balanced approach has helped us navigate challenges and continue innovating in the financial services industry.”
Evolution Of Saifr’s Technology
How has the company’s technology evolved since launching? Herard noted:
“While our tech stack has largely remained consistent since our inception, the deployment and utilization of AI has evolved significantly, as we have adopted technology advances over the past few years.”
“We use a combination of AI, incorporating machine learning, deep learning, and generative AI. As LLMs became more powerful, we started incorporating synthetic data into our training process, which not only diversified our data sources but also enhanced our models’ efficiency and predictive accuracy. We have built custom, fine-tuned LLMs specifically for certain use cases. We have also developed models with a cost-efficient layering of LLMs of a variety of sizes to solve specific problems. And, we are now working on ways to use agentic AI. We have an unbelievably smart AI team that can keep up with the advances in AI to harness this new technology to help solve our clients’ problems.”
Significant Milestones
What have been some of the company’s most significant milestones? Herard cited:
“I’m proud of the significant milestones we’ve achieved in the short time since our launch. We’ve consistently hit our internal Annual Recurring Revenue (ARR) targets, demonstrating our strong market traction. A major step in our growth was the acquisition of capabilities from Giant Oak, GOST, which we’ve since rebranded as SaifrScreen. With the acquisition, our solutions include capabilities for marketing compliance review and adverse media screening and monitoring.”
“In addition, it is always a milestone when you start getting industry recognition. Our recent placements on FinTech Global’s 2024 AIFinTech100 List and winning the “Best AI Solution for Regulatory Compliance Award” have been incredible honors. Additionally, being named “Most Innovative AI in Regulatory Compliance Initiative” by the A-Team Group Innovation Awards and “Best AI-based Solution for Financial Services” by the AI Breakthrough Awards further solidifies our position as a leader in the field. These achievements highlight Saifr’s rapid growth and our emerging role as a frontrunner in innovative regulatory technology for the financial sector.”
Differentiation From The Competition
What differentiates the company from its competition? Herard affirmed:
“What differentiates Saifr from our competition is our robust use of advanced AI in both content compliance and AML/KYC. For content compliance, our AI is uniquely sophisticated due to our proprietary data. This data informs our suite of content review products and enables compliant content to be in market much faster.”
“Our AML/KYC solution offers a unique layering of advanced AI, including specialized LLMs and NLP models, that work together to accurately determine risk matches. These extra layers of AI allow us to understand the nuances and sentiment of diverse data sources, such as unstructured data, resulting in more precise and comprehensive risk assessments.”
Future Company Goals
What are some of the company’s future company goals? Herard concluded:
“Our vision is to continue creating AI solutions that can tackle the complex challenges of regulatory compliance and business risks while maintaining the crucial balance of human oversight and expertise. We have several goals that come off that vision.”
“Some of our goals evolve around bringing our expertise to additional risk management challenges, which may include additional industries and/or geographies.”
“We have other goals around AI and other future technologies. For example, we are currently focused on expanding our AI capabilities to address more complex, multi-step problems in financial services compliance. While current AI applications excel at single-step tasks, we recognize that real-world challenges are rarely that straightforward.”
“Ultimately, we are trying to use advanced technology to help our clients be more accurate, effective, and efficient in their risk management to help protect them, their clients, and their industries.”
