Fast-Tracking Productivity Wins with Public AI Models

In this issue, we’re diving into practical ways to unlock productivity with publicly available AI models. We’ll cover the top tools, share actionable tips on implementing AI smoothly, and discuss recent news to keep you informed on the latest AI advancements.

If you’re ready to boost efficiency and leverage AI’s full potential, you’re in the right place!

The Balanced View: Instant Productivity Gains with Public AI Models

Public AI models are a natural choice for finance teams looking to enhance productivity without a lengthy implementation process. Accessible through straightforward subscriptions, tools like OpenAI’s GPT and Anthropic’s Claude can transform workflows almost immediately. These models deliver advanced capabilities for drafting reports, generating data insights, and handling time-consuming tasks, offering quick wins with minimal setup.

Why Public AI Models are the Practical Choice

Unlike custom or in-house AI solutions, public AI models don’t require complex integration or development time. Once subscribed, teams can start using these tools right away. Here’s why they’re especially practical for finance:

• Immediate Access and Setup: With a paid subscription, finance teams can be up and running within hours. There’s no need for lengthy installations or complex IT setups, making it an accessible first step into AI.

• Quick Productivity Gains: These models' capabilities—such as automated report generation, data summarization, and predictive analysis—deliver productivity boosts right from the start. Finance professionals can expect to spend less time on repetitive tasks, allowing more focus on high-value strategic work.

• Flexible Policies and Easy Training: Implementing a clear usage policy is straightforward and can be tailored to ensure that team members use the tool effectively and securely. With a few training sessions, users can learn to maximize the model’s potential while respecting data security protocols.

I was trying to find market data or use cases to demonstrate the speed of implementation and productivity gains, but I couldn't find anything meaningful. Instead, here are examples from my practice:

• Investor Report Automation:

  • Time spent before AI: 24 hours

  • Time spent after AI implementation: 4-5 hours

• Cash Flow Projection:

  • Time spent before: 3 hours per week

  • Time spent after: 30 minutes per week

• Multi-Entity Report Consolidation:

  • Time spent before: 30 hours per month

  • Time spent after: 6 hours per month

In all cases, the implementation took no more than 2 weeks.

How to Make Public AI Models Work for You

To quickly realize the benefits of public AI models, consider these practical steps:

• Get Your Subscription Set Up: To explore the model's capabilities, start with a standard Plus/Pro or Team subscription. Make sure to turn off the “Sharing data with everyone” setting.

• Establish an AI Policy: A clear policy can address essential aspects, such as what data can or cannot be entered into the model. This step doesn’t have to be complex, but boundaries should be set to keep usage aligned with company standards.

• Train the Team: Set up brief sessions on crafting effective prompts and practical applications, like drafting initial report templates, automating recurring communications, or performing data analysis summaries.

Realizing the Benefits in Weeks

With public AI models, finance teams don’t have to wait to see results. Common applications like drafting reports, summarizing data, or providing real-time analysis can remove hours of repetitive work every week. Teams can also use these tools to stay agile, handling month-end close tasks, investor updates, and ad hoc analysis in record time.

This approach enables teams to see returns on investment from day one, making AI adoption not just practical but impactful.

AI Tool Spotlight: The Overview of the Top AI Models

Here’s a look at the leading options I am using—OpenAI’s GPT and Anthropic’s Claude—with a focus on their practical applications and unique features for finance professionals.

1. ChatGPT (OpenAI)

• Overview: OpenAI’s GPT model excels in generating text, summarizing data, and answering questions with high accuracy. Team or Enterprise subscriptions offer enhanced security and customization options.

• Why It Works for Finance: OpenAI’s GPT is versatile enough for drafting and automating reports and generating financial narratives, such as quarterly overviews or investor updates. It works with various types of data and files (including Excel!) and can handle large data volumes.

• Especially Useful Features:

  • Web Browsing: ChatGPT’s browsing ability enables it to pull the latest financial data, industry news, or regulatory updates directly from the web.

  • Canvas Mode in GPT-4: This feature allows users to visually interact with the interface, providing on-screen tools to edit, annotate, and adjust text.

  • Custom GPTs: Tailored GPT models can be created for specific tasks, like auditing or forecasting, making the tool even more aligned with specialized workflows.

2. Claude (Anthropic)

• Overview: Claude by Anthropic emphasizes safety and interpretability, making it particularly attractive for teams concerned with data integrity. Claude is adept at handling nuanced language tasks and offers strong performance in reasoning and summarization.

• Why It Works for Finance: Claude is excellent at producing well-organized summaries, drafting communications, and assisting with data-driven insights. Its robust privacy and compliance features make it suitable for environments where data security is important.

• Especially Useful Features:

  • Artifacts: Claude’s Artifacts feature is highly versatile. It allows users to save, share, and manage various types of content, including text, tables, interactive dashboards, and diagrams.

  • Shared Workspaces: On the Team plan, Claude offers shared Project spaces, making it easier for teams to collaborate.

  • Improved Data Handling: Claude emphasizes secure data handling, ensuring that sensitive financial data is managed with high security and compliance.

Note: I’ve heard good things about Google Gemini, especially regarding its integration with Google Suite, but I haven’t personally used it. Microsoft Copilot seems to have mixed reviews—many users find it not quite ready for complex finance tasks yet.

Choosing the Right Model for Your Team

In my daily work, I regularly switch between ChatGPT and Claude, as each brings unique strengths. I appreciate Claude's Artifacts feature and Projects organization, which make it highly effective for managing and revisiting content. For drafting documents, I find ChatGPT’s 4o with Canvas model particularly useful, and its robust data processing capacity is great at handling large datasets.

A key differentiator is ChatGPT’s web browsing capability, which Claude currently lacks. This makes ChatGPT more versatile for gathering real-time information directly from the web.

I recommend trialing both tools for a month to see how each aligns with your needs before committing to any automation processes. If you need to commit to just one from the start, here are some guidelines:

• Choose ChatGPT for extensive data processing, such as uploading and analyzing Excel files or integrating AI with Google Workspace or OneDrive.

• Use Claude if your focus is on safety and privacy. Claude’s Artifacts feature is excellent for organizing content securely, and its privacy features are well-suited for sensitive data. You may complement Claude with a free Perplexity account for occasional AI-driven searches.

Ask the AI CFO

Q: What are the costs of implementing publicly available AI, and how can I determine the ROI?

A: Great question! Investing in an AI model is about balancing upfront costs with the productivity gains it can deliver. Let’s break down the typical costs and ROI considerations to help you make an informed decision:

Scenario Overview

To illustrate, let’s consider a small company that has implemented simple AI automation for two specific processes. Although these aren’t the most impactful areas, starting with relatively low-risk automation allows the team to gain confidence in AI, demonstrate measurable success, and refine their approach before expanding into more critical workflows.

The company targeted two highly manual, low-risk workflows: financial report preparation and expense report auditing.

Cost Breakdown:

• Claude Teams Subscription: $125 per month for 5 users = $1,500 annually

• Team Training (e.g., AI Blend Workshop): $3,000

• Training Time Cost: 5 team members x 4 hours = 20 hours, totaling $1,000

• Implementation Costs: 20 hours of internal effort x $50/hour = $1,000

• Consulting Support: 5 hours x $300/hour = $1,500

• AI Community Subscription (e.g. AI Finance Club): $1,000 annually

Total First-Year Cost: $1,500 (Claude) + $4,000 (training) + $2,500 (implementation) + $1,000 (AI Club) = $9,000

Productivity Gains and ROI Calculation

Financial Report Preparation

• Previous Time Spent: 48 hours per month

• Time After AI: Reduced to 10 hours per month

• Monthly Savings: 38 hours saved at $50/hour = $1,900 per month

• Annual Impact: $1,900/month x 12 months = $22,800 annually

Expense Report Auditing

• Previous Time Spent: 40 hours per month

• Time After AI: Reduced to 3 hours per month

• Monthly Savings: 37 hours saved at $50/hour = $1,850 per month

• Annual Impact: $1,850/month x 12 months = $22,200 annually

Total Monthly Savings: $1,900 (report prep) + $1,850 (audit) = $3,750 per month
Total Annual Savings: $45,000 annually

This implementation achieves:

• First-Year ROI: 400%

• Monthly Savings: $3,750 per month in time and labor costs

• Two-Year ROI: 850% cumulative

With two very simple automations, the team realized $3,750 in monthly savings and meaningful productivity gains from the start. As additional high-value AI workflows are identified, ROI potential will increase further, driving more substantial monthly and annual savings.

AI News of the Week: Desktop Versions of ChatGPT and Claude

ChatGPT and Claude now both have desktop apps, adding a new level of accessibility and functionality for users. These desktop versions are designed to improve system integration, provide faster response times, and include specialized tools for multitasking and file handling. Finance professionals, in particular, can leverage these updates for streamlined workflows, quick data analysis, and enhanced productivity—all without relying solely on web browsers.

• Quick Access with Keyboard Shortcuts: The Claude desktop app can be opened instantly with a keyboard shortcut (Ctrl + Alt + Space), ideal for finance professionals who need quick AI support while multitasking in fast-paced environments.

• Standalone Application: Both apps run independently of web browsers, reducing distractions and supporting a more focused workflow.

• Companion Window: ChatGPT’s desktop app offers a companion window that stays on top of other applications, allowing professionals to interact with ChatGPT seamlessly while working on spreadsheets or documents.

• File Integration: Desktop versions enable easy uploads of local files or screenshots, helping teams provide context-specific information for more accurate AI assistance.

If you frequently use ChatGPT and/or Claude, consider trying out the desktop versions. I have installed both!

Closing Thoughts

I hope this issue has inspired new ways to integrate AI into your daily work and demonstrated the potential of accessible AI tools. As AI technology advances, tools like ChatGPT and Claude become ever more capable allies for finance teams looking to enhance productivity. If you’re using these tools or trying out the new desktop versions, let me know your experience—we’d love to hear your feedback!

Stay tuned for more updates and practical insights in our next issue as we continue to explore the evolving world of AI in finance.

We Want Your Feedback!

This newsletter is for you, and I want to make it as valuable as possible. Please reply to this email with your questions, comments, or topics you'd like to see covered in future issues. Your input shapes my content!

Want to dive deeper into balanced AI adoption for your finance team? Or do you want to hire an AI-powered CFO? Book a consultation! 

Did you find this newsletter helpful? Forward it to a colleague who might benefit!

Until next Tuesday, keep balancing!

Anna Tiomina 
CFO & AI Enthusiast