Gradient Labs is an AI company that has built (what they claim to be) the only safe AI customer service agents for regulated industries, including financial services, fintech, and insurance.
While AI-powered customer service isn’t new, regulated industries have been left behind due to strict compliance requirements and the high cost of mistakes.
Gradient Labs claims that it changes that.
Their AI customer service agents reportedly achieve 80-90% resolution rates, 75% cost savings, and a 98% quality assurance pass rate (compared to 90% for humans), outperforming both human agents and traditional bots.
Gradient Labs’ CEO Dimitri Masin, former VP of Data Science, Financial Crime and Fraud at Monzo (UK’s Venmo Bank), has shared key insights with CI on the transformative impact of AI in regulated industries.
He has shared his perspective on the barriers to AI adoption in highly regulated industries and how companies can navigate compliance while leveraging automation;
He has also commented on how AI-driven automation is reshaping customer service, enhancing efficiency, reducing operational costs, and redefining workforce roles in financial and regulated sectors.
Dimitri Masin is the co-founder and CEO of Gradient Labs, the artificial intelligence startup redefining customer service in regulated industries.
With a background in financial engineering and AI, Dimitri specializes in safe and compliant automation.
Masin has previously held roles with Google and Monzo, where he served as the Vice President of Data Science, Financial Crime, and Fraud.
Our conversation with Dimitri Masin is shared below.
Crowdfund Insider: AI is being adopted rapidly across financial services. Where do you see the biggest areas of impact in the next 2-3 years?
Dimitri Masin: AI has been helping the financial sector for some time. However, we’ve moved beyond small efficiency gains, with the technology now capable of automating entire workflows and fundamentally reshaping operations.
In customer operations, for instance, AI isn’t just responding to inquiries and routing customers to the right resources. It’s involved in risk assessments, fraud detection, compliance screening, and other critical backend processes with remarkable accuracy.
Likewise, AI is taking on historically slow internal processes—handling document processing, regulatory compliance, and claims management—to speed up operations and reduce the menial workload.
Moving forward, I believe fraud and money laundering prevention will be the sector’s primary focus, given it cost institutions and their customers over $1 trillion last year. As financial crime grows more sophisticated, the industry must continue to shift away from rule-based fraud detection to AI-driven anomaly detection systems—machine learning models trained to recognize subtle clues and identify fraudulent activity that a human analyst would undoubtedly miss.
Crowdfund Insider: Some parts of financial services (fraud detection, trading) have embraced AI, while others (customer service, underwriting) lag behind. Why do you think some areas are slower to adopt AI?
Dimitri Masin: The cause of slow AI adoption is multi-faceted: firstly, you have the technological limitations, then the risk factor, regulatory complexity, and the sheer cost of getting it wrong.
Firms have been experimenting with AI in areas such as fraud detection and algorithmic trading for years because the benefits far outweigh the risks, and successful implementation can deliver immediate and substantial financial rewards. It also helps that these functions are data-driven, which makes them well-suited to AI development.
However, authentic, human-like customer service interactions were impossible until just a few years ago with the release of GPT 3.5, which marked a seismic shift in what is possible using generative AI.
Of course, rudimentary automation solutions have been available for several years. However, for highly regulated businesses, it was a question of whether the benefits were worth the effort and risk—Even a minor error can have considerable regulatory and reputational consequences. Financial institutions didn’t want to make the headlines or face a lawsuit because AI made an unfair lending decision or mishandled a vulnerable customer, so they stuck to what they knew worked.
Crowdfund Insider: However, as AI models become more advanced, transparent, auditable, and controllable, these barriers are starting to come down.
Dimitri Masin: Financial institutions are notoriously risk-averse when it comes to AI. What strategies should banks and fintechs use to overcome internal resistance to automation?
AI significantly outperforms humans at spotting vulnerabilities, identifying fraud, and preventing it—delivering faster, more accurate results at a lower cost. This benefits all stakeholders, whether businesses, customers, or regulators. However, the challenge lies in convincing decision-makers that these quality improvements justify the investment.
Our approach is to present clear, objective measures of quality and success and compare how human teams perform against AI based on the same criteria. It’s easy to claim that the technology is revolutionary as an AI company founder. However, the measure of its impact lies in the data, and financial service leaders ultimately want to see that the numbers add up.
Crowdfund Insider: Many financial institutions hesitate to fully automate because of concerns about public backlash and job losses. How should companies balance AI-driven efficiency with the social responsibility of employment?
Dimitri Masin: The fear of automation often stems from misunderstanding. AI won’t trigger a sudden wave of layoffs but rather a shift in how human teams work.
The first wave of AI adoption will automate repetitive, low-value tasks, such as account verifications, answering common questions, and transaction monitoring. Essentially, we’re putting AI in charge of the monotonous, burnout-inducing chores so that the workforce can focus on tasks that require creativity, skill, and deep thinking—complex cases, personalized support, and strategic problem-solving, which deliver the most value and we actually enjoy doing.
At the same time, automation is drastically improving customer experiences, delivering seamless and efficient services that consumers no longer appreciate but expect. As we’ve seen countless times, institutions that fail to keep pace with evolving demands rarely survive, which is a far greater risk to the workforce.
That said, businesses need to handle adoption responsibly. Automation is inevitable, but how it’s implemented will determine its impact on the job market. If we want to empower workers rather than simply replace them, upskilling is a must.
Crowdfund Insider: Historically, automation has displaced jobs but created new ones. Do you think AI in fintech will follow this pattern, or is this time different?
Dimitri Masin: History has shown time and time again that as needs change, new roles emerge. However, this time feels different—we are entering a new era where humans are no longer required to perform the majority of work.
Within the AI community, the established stance is that concerns over job displacement are overstated and that there’s little cause for alarm. It may be an attempt to avoid panic or wishful thinking, perhaps. However, I believe AI will eliminate far more jobs than it creates. It’s an uncomfortable prospect, but ignoring it doesn’t make it any less true.
The sooner we—across all industries, not just fintech—acknowledge the problem, the sooner we can begin discussing and devising solutions to ensure a sustainable future.
Crowdfund Insider: With rising interest rates and economic uncertainty, many fintechs are under pressure to cut costs and become profitable. Will AI-driven automation be the key to fintech survival?
Dimitri Masin: OpenAI’s Sam Altman recently coined the term “one-person unicorn”—a billion-dollar company operated by a single person. With the rapid pace of innovation in the AI space, this could become a reality within the next five to ten years.
If that’s the case, how can companies with thousands of employees possibly compete? To survive, it’s critical that businesses across all sectors—not just fintech—integrate AI into their operations and workforce.
Crowdfund Insider: Regulators worldwide are taking a closer look at AI-driven decision-making. Do you think we’ll see major AI-specific regulations in financial services, similar to GDPR for data privacy?
Dimitri Masin: AI regulation in financial services is fragmented and reactive, with clear guidelines long overdue.
As regulators catch up, I expect an AI governance framework similar to GDPR to materialize. With a likely focus on transparency, accountability, and bias mitigation, financial institutions will need to prove their AI systems are fair, auditable, and explainable.
Hopefully, decision-makers will be able to find the right balance—avoiding overly restrictive policies that slow progress, while maintaining oversight and minimizing risks.
Crowdfund Insider: If you had to make one bold prediction about AI and financial services for 2030, what would it be?
Dimitri Masin: Most financial institutions will be AI-first companies by 2030, and AI won’t just be another tool in the stack. It will be the foundation of how banks, fintechs, and insurers operate.
The companies that embrace AI today will lead the future of financial services. Those that don’t? They will be left behind and consigned to the history books.