Hey Reader,

"Is it safe to use AI in finance with all the data and rules we have?"

It's a good question.

After all, one misused AI tool could show private data, cause problems in audits, or give bad advice.

In a 2024 report by Deloitte, 48% of CFOs cited "Generative AI adoption" as the internal risk that worries them the most.

Here's the problem though:

Yes, AI has some specific security and compliance challenges because of how much data it uses, the decisions it makes on its own, and how it "learns" over time.

• Keeping data private and controlling access
• Understanding how AI makes decisions
• Making sure outputs are reliable and consistent

...those are all valid concerns.

But these risks get much easier to handle once you understand them.

More importantly, not using AI at all is an even bigger risk.

But here's something that I think you must attend to...

Future of Finance with AI (free webinar)

🗓️ Date: Tuesday, April 1
​
Join me, alongside Nilus CEO Daniel Kalish and treasury expert Matt Harlan, for an interactive, eye-opening webinar on how AI is revolutionizing finance.

Here's what you'll learn:

• What's really holding finance teams back today (and how to break through)
• Real-life examples of AI transforming treasury, cash management, and accounting
• Practical insights on overcoming common barriers to AI adoption in your company
• A clear vision of how AI will define the next era of finance

Don't miss this chance to move beyond buzzwords and see AI's powerful impact firsthand.

​📌 Spaces are limited—Reserve your spot now!​

The Cost of Fearing AI is High

By avoiding AI because of misplaced security fears, you're:

1. Wasting Time on Manual Work. Every hour spent on tasks AI could automate is an hour NOT spent on strategy, analysis, and high-value work. You're leaving huge productivity gains on the table.
2. Using Outdated Insights. The longer you wait to use AI's predictions, the longer you base decisions on old data. You could be the first to see new opportunities or risks.
3. Limiting Your Team's Skills. As AI becomes normal in finance, teams that don't build hands-on experience now will struggle to catch up later. You're risking your team's relevance.
4. Stifling Innovation and Competitiveness. With 58% of finance teams using AI in 2024 and 21% more planning on using AI soon (Gartner), those that don't will fall behind on efficiency and insights. The opportunity cost compounds massively over time.

And don't worry - AI won't replace you anytime soon.

At least, Gartner doesn't think so. One of their September 2024 press releases stated:

"By 2026, 90% of finance functions will deploy at least one AI-enabled technology solution, but less than 10% of functions will see headcount reductions."

So don't be scared of AI.

You Can Use AI Safely (If You Know THIS)

Here's the good news - you DON'T have to choose between AI's benefits and tight security/compliance.

With the right approach, you can have both.

The key is to stick with well-known AI providers.

The big firms have strong security (encryption, access controls, etc.) and compliance certifications (SOC2, ISO27001, etc.). Most have the same or higher standards as your existing cloud tools.

• If you're using Google Drive, you might as well use Google Gemini.
• If you're using Outlook in the cloud, you might as well use Microsoft Copilot.
• If you're using Amazon Web Services, you might as well use Amazon AI services.

Of course, you always need to check that your company has approved the AI tools of these three giants. However, all of them are really clear that the security standards are the same for their cloud and their AI tools.

"Can I use ChatGPT safely?"

If you use ChatGPT Enterprise, it's designed for business security and compliance:

According to OpenAI's documentation, your data and conversations:

• Are NOT used to train OpenAI models
• Are protected by SOC 2 compliance
• Are encrypted both in transit (TLS 1.2+) and at rest (AES-256)
• Are controlled through enterprise features like: • SSO authentication • Domain verification • Admin console for managing access • Usage analytics and monitoring

As OpenAI states, "You own and control your business data in ChatGPT Enterprise." This means your company maintains ownership of both inputs and outputs while getting the security controls needed for enterprise use.

So if you're dealing with sensitive data, use ChatGPT Teams or Enterprise.

If you're not dealing with mega-sensitive data, you can use the normal ChatGPT website. You can get far by anonymizing datasets before feeding them into AI. More on this in a future lesson.

“How about other tools?”

If you need more tools than that, a simple Google search is enough to tell which AI services have enterprise-level security.

For sensitive business data, use enterprise versions:

• ChatGPT Enterprise
• Microsoft Copilot Business
• Google Gemini Enterprise
• Anthropic Claude (Enterprise version)

These enterprise versions typically offer:

• SOC 2 compliance
• Data encryption
• Access controls
• Usage monitoring
• No training on your data

For non-sensitive data, you can use standard versions but should:

• Avoid sharing confidential information
• Be careful with personal data
• Consider anonymizing data where possible
• Check your company's AI usage policies

"How do I know if an AI tool is enterprise-ready?"

Look for:

1. Security certifications (SOC 2, ISO 27001)
2. Enterprise features (SSO, admin controls)
3. Clear data privacy commitments (like GPDR in Europe)
4. No training on customer data

The key is matching the tool's security level to your data's sensitivity. When in doubt, consult your IT security team.

Crawl, Walk, Then Run to AI Success

You don't need to dive into the deep end of AI on day one.

The smartest approach is to start small, learn the fundamentals, and scale up carefully. Here's how:

Phase 1: Safe Experiments

• Try low-risk internal use cases (e.g. drafting reports)
• Build AI knowledge and prove value across your team
• Develop initial control frameworks and best practices

Phase 2: Controlled Pilots

• Pilot-specific AI uses with tight oversight
• Rigorously check security, outputs, and compliance
• Measure efficiency gains, cost savings, and decision value

Phase 3: Strategic Scaling

• Expand AI to more core finance based on pilots
• Formalize policies, success metrics, and center of excellence
• Promote AI value to partners for wide adoption and ROI

With this phased approach, you can get AI's rewards NOW while managing its risks at every step.

Nicolas
​

P.S. Don't forget to join me for my next webinar on the Future of Finance with AI on 1st April:
​📌 Spaces are limited—Reserve your spot now!​