From Intrigue to Implementation: 3 Phases of AI Integration
This three-step approach can move your accounting team’s AI curiosity to action with real results.
By Twyla Verhelst, CPA | Digital Exclusive – 2026
Despite constant headlines, most accounting firms are still approaching artificial intelligence (AI) with cautious curiosity rather than with real commitment. This hesitation is reflected in recent data.
According to Karbon’s 2026 “State of AI in Accounting” report, 83% of accounting professionals say they’re excited or intrigued by AI’s potential, but only 18% of respondents thought their peers shared this enthusiasm.
This disconnect points to a larger problem across accounting firms: They’ve been thinking about AI far more than they’ve been doing anything meaningful with it.
As AI’s influence continues to expand, however, more accounting leaders and professionals are starting to ask how they can actually use AI without breaking everything.
Fortunately, making the transition from intrigue to implementation doesn’t have to be as tall a task as it seems—successful AI integration starts with these three phases.
1. Individual Exploration
Individual exploration is where AI adoption begins—where comfort and confidence are born. Here are some examples of how you can begin to explore AI’s possibilities:
- Curiosity: Try entering different prompts into ChatGPT, Google Gemini, or Microsoft Copilot to see how it responds.
- Experimentation: Test AI tools on specific tasks, such as drafting emails or summarizing documents.
- Integration: Embed AI into your workflows, such as using an internal note-taker and building repeatable checklists for ongoing work.
- Leading: Take an AI-first approach to shape your operations around technology. This might include building formal systems or hiring someone to specifically manage your AI processes.
It’s important to note that this first exploratory phase is likely to exist at different levels for different people in your firm. Wherever someone falls on this spectrum, adaptability—the willingness to lean into the possibilities of AI with an open mind and a desire to experiment—matters more than current skill level.
For example, maybe one of your team members is testing prompts in ChatGPT. While some may see this as a trivial exercise, it’s far from it. It’s an easy, low-risk way for a person to become more familiar with AI, gain confidence in their skills, and increase their proficiency with a tool. They’re climbing the sometimes-steep learning curve of AI and seeing what it might help them do faster or better than before. Additionally, this is someone who’s likely to become more open to AI experimentation in other areas—and having a team of people like that is a great way to start making meaningful AI strides in your firm.
2. Team Integration
This is the phase where collaborative AI use and knowledge sharing enter the chat. This is when AI experimentation in a firm moves from just a handful of people exploring tools individually to bringing the full team along on a collective journey.
Team integration can also be a challenging phase. I’ve heard from plenty of pro-AI firm leaders who paid for a premium ChatGPT subscription or some other powerful tool for their staff only to find out that no one was using it. This is a frustrating and familiar situation, but it’s also not terribly surprising. If you want to get people excited about AI, you can’t just roll out a new tool and hope they make the most of it. The goal instead should be a proactive approach that encourages collective experimentation. Here are some ideas that can help:
- Hold weekly AI roundtables: Provide time for your team to share successful discoveries and applications, as well as things that didn’t work.
- Present a prompt of the week award: Highlight effective AI prompts and solutions that addressed real problems in the firm.
- Create an AI Slack or Teams channel: This can serve as a repository of proven prompts for common tasks and resources with helpful guidance.
- Assign rotating AI champions: Designate team members on a rotating basis to lead AI discussions and share learning strategies.
3. Operationalizing AI
This last phase is where AI gets a little more intentional—it’s where you look for ways to move from ad hoc experiments to systematic AI implementation in your firm’s processes. Specifically, this phase involves three big shifts:
- Across workflows: AI is used throughout the firm’s processes and procedures, not just for one-off tasks.
- Multiple tools: AI starts to become part of multiple areas of the firm, such as your practice management tools, tax software, or meeting transcription tools.
- Automations: AI is used to connect tools and data to trigger actions in the firm, not just to produce content that can be cut and pasted.
Some specific examples in this operational phase may include using AI to perform financial analysis, transcribe meetings, or streamline client onboarding. More advanced use cases might include building a custom ChatGPT tool to draft client emails using the firm’s voice and tone.
Overall, while this three-phase approach to AI integration can serve as a helpful guide, how every firm gets there will be different. As you undertake this journey, it’s important to meet your team where they are. If you jump too fast from one phase to the next when your team isn’t ready, then your AI adventure will likely be a short and disastrous one. That’s why it’s important to evaluate where your team falls on the AI spectrum right now. Are they excited, scared, or clueless? Finding the sweet spot starts with looking at your firm and asking one key question: “What’s one thing that feels harder than it should, and how can AI help us reimagine it?” Coming up with that answer together as a team is a great way to move the needle from AI intrigue to AI integration.
Twyla Verhelst, CPA, is vice president of industry relations and community at Karbon.
Related Content: