How AI is transforming compliance

As AI reshapes the private markets, most of the attention has gone to front office applications—deal sourcing, LP communications, and fundraising. However, AI is also being used to transform one of a firm's most foundational functions—compliance.

We sat down with Carlo di Florio, President of ACA Group, which provides AI enabled compliance technology platform and modules, to explore how firms are already using AI in their compliance operations. Here are some of the most compelling use cases:

di Florio noted that the typical responsibilities of a compliance officer—what he called the “front end” of compliance—include:

• Inventorying the regulations that apply to the business

• Conducting a risk assessment to rank regulatory exposure

• Developing and updating policies and procedures

• Communicating and training on those policies

• Testing whether policies are being followed

• Documenting and preparing for regulatory exams

• Investigating and remediating any red flags

And while there's already technology that automates many of these processes, he explained that AI brings new power to this existing tech stack by:

• Identifying regulatory changes and suggesting policy updates in real time

• Supporting dynamic testing protocols that adapt based on new risks

• Streamlining documentation and readiness for regulatory exams

AI is increasingly used to support internal compliance around personal trading, political contributions, and outside business activities.

Firms are using AI tools to:

• Pre-clear trades against firm holdings

• Monitor political donations to stay within regulatory bounds

• Flag potential conflicts of interest tied to external affiliations

Even regulators are using AI to ensure a more balanced market. Germany's financial regulator, BaFin, has implemented AI in its market analysis systems to detect market abuse and suspicious trading patterns more effectively.

di Florio noted that marketing content is one of the most heavily regulated areas in financial services, with both FINRA and SEC rules governing what firms can and can’t say in promotional materials.

ACA reviews over a million pieces of marketing material per year across its client base, making it one of the largest reviewers in the U.S. Historically, this review process has been manual and resource-intensive, relying on teams of experts. To streamline this work, ACA has developed a marketing review technology platform that integrates AI to assist with reviewing content.

The AI is trained on rule sets and historical data to:

• Review marketing materials

• Flag language that may not be fair, balanced, or complete

• Recommend alternative phrasing that would be more compliant

The AI operates with “human in the loop” oversight, meaning that compliance experts still review and validate AI-suggested changes before anything is approved or finalized.

Analysts and portfolio managers often consult external experts to inform their investment theses. These conversations must be monitored to ensure that no material, nonpublic information (MNPI) is shared that could illegally influence investment decisions.

Historically, expert network chaperoning required a compliance officer sitting on the call and taking notes. AI is now changing that by:

• Joining and transcribing calls in real time

• Flagging references to MNPI

• Creating structured reports for Chief Compliance Officers

In some firms, the transcripts are retained as research. Others choose to retain only the risk flags, an example of AI being tailored to a firm's risk appetite.

As di Florio shared, AI provides ACA with a scalable, consistent, and auditable way to oversee these interactions without relying solely on manual effort. “It's one of the best use cases we've seen for compliance AI,” he said.

Electronic communication surveillance (ECS) refers to the monitoring, reviewing, and analyzing of digital communications, including emails, instant messages, video calls, social media interactions, and texts. In regulated industries like financial services, e-comm surveillance is required by law to ensure employees aren't engaging in illegal activities such as insider trading, market manipulation, or violating conduct rules.

AI is transforming ECS by making it smarter, faster, and more precise. Traditionally, ECS systems relied on static keyword lists and manual reviews, leading to an overwhelming number of false positives and missed signals. With AI, firms can now monitor communications at scale while surfacing genuinely risky behavior, often in real time. ACA, for instance, utilizes tools like natural language processing (NLP) to monitor dozens of platforms including Teams, Zoom, and WhatsApp in more than twenty five languages

A report from Global Relay found that 31% of survey respondents are already using AI for communications surveillance (or plan to do so in the next 12 months). BNY Mellon, for instance, is using AI to analyze structured and unstructured data across multiple communication channels in over 16 languages, enabling proactive detection of compliance risks and strengthening the firm's overall risk management framework.

Each year, an estimated $800 billion to $2 trillion is laundered globally, according to the United Nations Office on Drugs and Crime (UNODC)—a staggering scale that underscores the importance of robust AML and KYC controls.

With more than twenty seven million names screened in 2023 for KYC/CIP requirements, ACA understands better than most how labor-intensive AML/KYC can be. AI solutions can dramatically increase the speed and accuracy of:

• Identity verification

• Document validation

• Risk scoring and anomaly detection

AI also enables real-time monitoring of customer profiles, also known as Perpetual KYC. This shift from periodic reviews to continuous oversight enhances compliance and reduces the risk of financial crimes.

It's important to note that AI doesn't absolve firms of responsibility. di Florio pointed out, "You can outsource activity, but you can't outsource responsibility."

But what's even more powerful than any individual use case? Interoperability.

As di Florio put it, "AI is most powerful on a platform."

When different systems talk to one another, AI can synthesize signals across domains. For example, a flagged personal trade can be cross-checked against firm transactions, email patterns, and recent marketing activity. That level of connected insight isn't possible without a unified platform, and that's where AI truly shines. As workflows become more connected, data becomes more structured, and platforms become more intelligent, compliance becomes better, faster, and more reliable.

AI is only as powerful as the platform behind it. That's why we built JunieAI—the first enterprise-grade AI designed specifically for private markets GPs. Explore JunieAI →