By Chris Jenkins, CAE | SC.CPA CEO
Key Takeaways
- AI’s biggest challenge is no longer capability. It’s economics.
- SC.CPA now processes more than 10 billion AI tokens every month, giving us a unique perspective on AI at scale.
- The best AI strategy isn’t using the biggest model for every task. It’s using the right model for the job.
- The firms that succeed won’t necessarily use the most AI. They’ll use it more intentionally by balancing performance, security, governance, and cost.
Artificial intelligence is entering a new phase of its rapid evolution. Just months ago, the challenge among organizations was who was adopting AI and how quickly. Now, the challenge is around effective use. Why the shift?
Today, SC.CPA processes more than 10 billion AI tokens every month across our internal workflows, automation, and AI tools. That kind of volume teaches you a few things pretty quickly. Token use is a key indicator of the power necessary to run AI-driven workflows. Everything from email generation to report summarization, moving up to complex workflows and research, has a token cost–the currency of AI. The heavier the AI lift, the more tokens used.
With that in mind, the biggest challenge we face with AI isn’t capability. It’s economics.
AI models keep getting better, and the leading companies like Open AI, Anthropic, and Microsoft are delivering models with varying capabilities. For instance, ChatGPT currently offers “Instant,” “Thinking,” and “Pro.” But we’re learning that throwing the most powerful model at every problem isn’t always the most economical move. As AI becomes part of everyday work, firms need to think beyond prompts and productivity gains. They need an approach that balances performance, cost, security, and long-term flexibility. We need to develop a habit of matching the model to the task.
The End of “Unlimited” AI
For the past several years, the AI industry has focused on rapid adoption. Models became more capable, prices kept falling, and many organizations got used to AI feeling almost unlimited.
Model providers are investing hundreds of billions of dollars in infrastructure while competing to deliver better performance at lower cost. At the same time, organizations are integrating AI into more of their operations, driving token usage and costs higher every month.
This presents an opportunity for organizations to actively manage AI consumption and manage the associated costs.
Five Lessons We’ve Learned
After integrating AI into nearly every part of our daily operations, five lessons have become clear.
1. Not Every Task Needs the Smartest Model
Sending every request to the most advanced AI model is like asking a senior partner to organize digital files.
Routine work (i.e. classification, extraction, summarization, formatting, document preparation, and workflow automation) often doesn’t require the largest or most expensive model available.
One of our earliest discoveries was that many routine workflows produced nearly identical results using smaller, more efficient models. That allows us to reserve our most advanced AI models for complex reasoning, research, and strategic work and dramatically improve cost efficiency.
Matching the model to the task delivers better long-term results.
2. Context Often Matters More Than Raw Intelligence
A model that understands your organization’s workflows, terminology, policies, and institutional knowledge will often outperform a larger model starting from scratch.
Building organizational context is becoming a competitive advantage.
3. AI Costs Deserve Active Management
Software has budgets. Cloud infrastructure has budgets. Cybersecurity has budgets. AI should too.
Organizations that ignore AI consumption today may be surprised by the cumulative cost of thousands, or even millions, of daily AI interactions across their workforce.
Good governance isn’t about limiting innovation. It’s about making innovation sustainable.
4. Hybrid AI Is Becoming the Practical Choice
One of our biggest lessons has been realizing that AI isn’t an either-or decision.
For many routine business processes, open-source models running locally perform remarkably well while keeping sensitive information inside the organization’s environment when appropriate. For complex reasoning, research, strategic planning, and creative work, frontier cloud models continue to deliver exceptional value.
5. AI Amplifies Expertise
The greatest gains we’ve seen haven’t come from reducing headcount.
We have eliminated repetitive work so professionals can spend more time delivering value to the membership experience. AI accelerates the work. Our team remains responsible for its judgment and authenticity.
In accounting, ethics, professional standards, and client relationships remain the cornerstone of trust and authenticity. AI can streamline the work that distracts you from showcasing your expertise.
Why This Matters for the Accounting Profession
CPAs work under extremely unique circumstances:
- You work with highly sensitive financial information.
- You operate in a regulated environment.
- You have an obligation to protect client confidentiality while continually improving productivity.
That requires a more thoughtful AI strategy than simply buying another subscription.
The firms that thrive over the next decade won’t necessarily be the ones using the most AI. They’ll be the ones who understand how to strategically use AI responsibly, securely, and economically.
Why We’re Sharing What We’ve Learned
Everything we’ve learned has come from real-world implementation, not theory.
- We’ve tested models.
- We’ve built workflows.
- We’ve experimented with local AI.
- We’ve measured costs.
- We’ve refined our governance.
- And we’ve learned plenty of lessons along the way.
What surprised us most was discovering that accounting professionals are asking the same questions we asked.
- How do we choose the right models?
- How do we protect sensitive information?
- How do we manage costs as AI usage grows?
- How do we build an AI strategy that’s sustainable instead of chasing the latest trend?
Those conversations inspired us to continue SC.CPA AI Mastermind for a second year.
The AI Mastermind isn’t about product demos or chasing the latest AI announcement. It’s about helping accounting professionals build practical AI strategies they can actually implement.
Together, we’ll explore topics including:
- Developing an AI strategy for your firm or organization.
- Choosing the right AI models for different business tasks.
- Building secure and responsible AI workflows.
- Managing AI costs as adoption grows.
- Understanding emerging AI tools and technologies.
- Learning from peers who are actively implementing AI in the accounting profession.
Whether you’re just getting started or already using AI every day, our goal is simple: help you make better decisions, avoid expensive mistakes, and build an AI strategy that creates lasting value.
The Next Competitive Advantage
The conversation around AI is changing. Long-term success won’t belong to the people who simply adopt every new AI model. It will belong to those who know how to combine technology, governance, economics, and professional expertise into a sustainable strategy.
If you’re ready to move beyond AI experimentation and build an AI strategy that’s practical, secure, and economically sustainable, we invite you to join the SC.CPA AI Mastermind.
- Learn from peers.
- Explore proven workflows.
- Build an AI strategy designed for the long term.
Because just subscribing to an AI to write your emails isn’t enough to have the competitive advantage anymore. Knowing when to use it, how to use it, and how to manage it is how the next generation of successful firms will separate themselves.
