Founded in 2020, Boston-based Saifr uses advanced AI technology to efficiently and effectively identify possible compliance and regulatory risks in content creation, adverse media screening, and e-communications for financial services and other firms. Saifr’s products help clients save time, reduce costs, and improve accuracy while protecting the business.
According to Vall Herard, CEO of Saifr, 2024 was a banner year for the company, successfully hitting key milestones of an acquisition and revenue increases.
In addition, he detailed, “We also expanded our client base beyond financial services into the marketplace industry/segment; and last but not least, we completed a strategic collaboration with Microsoft, where their CEO announced Saifr as one of the firms they are partnering with on industry models.”
Despite the powerful year, Herard explained that it wasn’t all smooth sailing, with the firm facing some challenges. While acquisitions are always hard, he claims, the integration of the acquired and acquirer can be much harder due to the integration of cultures.
“That being said, I’m really happy with how the Saifr leadership team coalesced and fully integrated the team that came on board in record time. So, while there are always challenges that are unique each time you do an acquisition, I think the fact that we were able to fully integrate the business in a little less than 10 months into our operations, is something to be proud of,” he stated.
AI in RegTech
The role of AI within financial services is evolving day-by-day, as new AI technologies come onto the scene. Within the RegTech space, AI is playing a huge role in supporting compliance.
Herard stated, “AI will play a big role in compliance for the simple reason that compliance is a highly manual requirement in regulated industries. One thing that AI is really good at is taking a process, synthesising it to remove a lot of the friction brought on by its manual nature, and making it more streamlined, efficient, and less costly.”
The Saifr CEO believes that AI will continue to play a large role in helping financial organisations meet their regulatory requirements and act as a ‘source multiplier’ helping firms be more efficient at remaining compliant.
One of the key challenges to date for businesses adopting and building AI models has been the issues surrounding safety and privacy, and making AI safer is a mission of high importance.
How can AI make AI safer? For Herard, Saifr’s mission is to make AI safer by embedding regulatory rules into AI models. “At its core, Saifr takes regulations, and turns them into executable code that can be used inside of an AI algorithm. By so doing, people can then use our models to make AI safer by ensuring that their AI applications comply with industry rules.”
Herard remarked that GenAI, which creates content, doesn’t know the regulations and may write that an investment is “the best,” include a money tree picture, or not know that footnotes are required. Saifr, he states, can act as a guardrail with Gen AI systems to help ensure that firms follow financial services regulations.
In his view, there is no greater manifestation of that than the fact the EU AI Act – a prominent European legislation – calls for AI models to be compliant with existing rules. “Saifr helps firms’ generative AI systems comply,” he said.
Herard explained that Saifr is also using AI for adverse media screening to help financial firms with their AML/KYC programs as well as to help trust and safety officers at marketplace companies. “In those two instances, we are helping businesses keep bad actors off their platforms, which makes these platforms safer for the regular consumer,” he said.
Herard went on, “We’re also using AI to help companies monitor communications – both marketing communications and electronic communications – to help ensure that employees, contractors and also consumers that those employees and associates are communicating with are staying compliant with the rules that are designed to make the industry safer.”
AI build-vs-buy
How should businesses think about build vs buy when it comes to AI? In the opinion of Herard, the trade-off between build vs buy has always been a debate, but history can provide some insights.
“There are lessons to be learned from previous technology cycles,” said Herard. “I don’t see AI as being any different in the sense that what we’ve seen in the past is that for medium size and smaller institutions, it is almost always preferable to buy vs build. The return on investment, the time to production, and the time to value can be greatly reduced for smaller or medium sized firms if they buy.”
To Herard, for larger institutions with more resources, that is not always a clear-cut case. “There are instances where, given the time to value that you’re trying to realise, it is easier and better to buy vs build. However, there are sometimes customisations that are so specific to a business, and where a business has the resources, where the build decision makes the most sense. So, it’s more a case-by-case basis for larger firms.”
The evolution of AI
In the view of Herard, one of the key developments for AI in 2025 will be the idea of model-as-a-service.
He explained, “I think this will be the dominant theme in 2025. I think model-as-a-service can lower the cost of building AI applications by at least 10x; and by so doing, lower the barrier to entry.”
In the case of Saifr, Herard outlined that the business is providing model-as-a-service so that firms can embed regulatory guard rails into the applications that they are building. This is key to Saifr’s collaboration with Microsoft, which is the idea that someone building an AI application can take Saifr models and embed them into their application – a trend Herard believes will continue in 2025 and beyond.
Future plans
As Saifr looks toward the future, what is next for the business? According to Herard, growth is a key watchword for Saifr. After experiencing a successful 2024, it’s all steam ahead for the company in 2025.
He added, “One of the things that we are really looking forward to is building on our model-as-a-service offering and releasing additional models into the Microsoft Azure AI Foundry. We’re looking forward to making some announcements in 2025 on that front.”
Copyright © 2025 RegTech Analyst
Copyright © 2018 RegTech Analyst
