Original post in TSCPA Journal
Artificial intelligence (AI) is no longer a theoretical conversation for the accounting profession. It is a practical, revenue-impacting reality. The problem firms face today is not a lack of interest in AI, but a growing gap between rising demand, constrained talent, compressed margins and client expectations that continue to escalate.
As firms move through another demanding tax season while navigating staffing constraints, margin pressure and rising client expectations, AI has become both a source of opportunity and concern. The questions we hear most from TSCPA members are not about whether to use AI, but how to use it responsibly, accurately and profitably.
This article addresses the most common themes raised by practitioners today, with a clear objective: helping firms move from curiosity and caution to confident adoption that drives client value, and ultimately, revenue growth.
Start With the Problem: Capacity, Risk and the Fear of Getting It Wrong
A consistent theme across recent industry coverage—from Inside Public Accounting to Accounting Today—is that AI adoption in accounting is being driven less by curiosity and more by pressure. Firms are being asked to do more work, faster, with fewer people, while maintaining or improving quality.
Within that context, one concern rises to the top:
“AI Gets It Wrong 50% of the Time”—Addressing Accuracy and Client Trust
Concerns about AI-generated answers are especially heightened during tax season or a complex audit, when the cost of an incorrect answer can be significant. Headlines and anecdotes suggesting that “AI is wrong half the time” have understandably created skepticism among both practitioners and clients.
The nuance here is critical. General-purpose, free, public AI tools are not designed to be authoritative sources for tax or accounting guidance. They are trained on broad datasets and lack the context, citations and guardrails required for professional use. When firms rely on these tools without appropriate controls or without supplying the necessary context, the risk of inaccurate output increases.
By contrast, domain-specific AI platforms—such as TaxGPT or CCH AnswerConnect.ai—are built on vetted, continuously updated tax and accounting content. At LBMC, we have taken this approach further by developing “LBMC Think”, a context-enriched AI platform designed to support research by drawing from a deep and continuously curated body of tax- and audit-specific knowledge.
This distinction has been emphasized repeatedly in recent Inside Public Accounting and Accounting Today coverage. AI risk is not inherent to the technology, but to how and where it is applied. These tools use AI to accelerate research, surface relevant guidance, and improve workflow efficiency, while still grounding outputs in authoritative sources. The AI does not replace professional judgment; it augments it.
How to address this with clients:
- Be transparent about how AI is used internally—as a research accelerator, not a decision-maker.
- Reinforce that all client-facing advice is reviewed and validated by credentialed professionals.
- Position AI as a quality enhancer: faster turnaround, more thorough research, and improved consistency.
Firms that proactively educate clients on this distinction build trust and differentiate themselves as modern, disciplined advisors.
The Expanding AI Tool Landscape: Moving Beyond the Hype Cycle
It seems every week brings a new AI tool, partnership, or platform announcement. Recent examples include the AICPA’s introduction of Josi for A&A professionals or the introduction of BlueJ, which offers AI-powered tax research. Rather than chasing “the latest and greatest,” firms should evaluate tools through a strategic lens:
- Is the tool purpose-built for accounting, tax, audit, or advisory?
- Does it rely on authoritative, citable content?
- Does it integrate into existing workflows and systems?
- Does it improve realization, margins, or client experience?
Beyond tax research and audit support, firms should also be aware of AI applications in the following processes:
- Revenue cycle optimization and billing analytics
- Client communications and knowledge management
- Back-office automation
- Contract, MSA, and engagement letter management
The firms seeing the greatest ROI are not those experimenting with the most tools, but those aligning AI investments to clear business outcomes.
From Curiosity to Capability: Best Practices for Implementing AI in Your Firm
One of the most common requests we hear, reinforced in both practitioner forums and national media, is for step-by-step guidance on implementing AI in practice.
While there is no one-size-fits-all roadmap, successful firms tend to follow a similar progression:
- Start with a business problem, not a tool. Identify where time is lost, margins are thin, or clients are underserved.
- Educate leadership and core teams. Shared understanding reduces fear and accelerates adoption.
- Pilot in controlled environments. Use AI internally before expanding its role.
- Establish governance and usage standards. Define where AI can and cannot be used.
- Measure outcomes. Track time saved, error reduction, realization and client satisfaction.
- Executive led. AI should not be treated as a technology ambition, but as an enterprise transformation goal. Organizations in which the CEO sets the direction for AI and empowers IT and transformation teams to execute that vision will be best positioned to lead in the AI era.
Prompting discipline is a key capability in this process. The Journal of Accountancy article on “Writing an Effective AI Prompt for an Audit” and the widely discussed “STAR” method (Situation, Task, Action, Result) provide practical frameworks for getting more reliable outputs from AI tools. Firms that invest in prompt training see faster adoption and more consistent results across teams.
Generative AI vs. Agentic AI: Why This Distinction Matters to Your Firm
For many practitioners, AI terminology itself can be a barrier. Two terms that frequently cause confusion are Generative AI and Agentic AI.
- Generative AI focuses on creating content including drafting memos, summarizing guidance, generating checklists, or explaining complex topics. It responds to prompts but does not act independently. Gartner will refer to this as ‘Everday AI’ and is best represented in the general-purpose, broad-based AI tools like ChatGPT, Co-Pilot and Gemini.
- Agentic AI goes a step further by executing multi-step processes within defined parameters such as extracting data, triggering workflows, or updating systems while still operating under human oversight. This often involves multiple specialized agents, or an orchestrating agent, working together to complete a task end to end.
In an accounting context, generative AI supports knowledge work while agentic AI has the potential to transform processes. This distinction was a central theme in recent Accounting Today discussions around AI and employment: the technology is not eliminating the profession but reshaping where human judgment is most valuable. Understanding the distinction helps firms adopt AI safely and incrementally.
AI, the Revenue Cycle, and the Question Everyone Is Asking: Will AI Take Jobs?
As I shared in recent Accounting Today coverage, AI should be viewed as a force multiplier. It enhances human expertise rather than replacing it. The firms that struggle are not those adopting AI, but those attempting to preserve legacy delivery models in a market that has fundamentally changed. Firms that align AI initiatives with revenue strategy, client experience, and talent sustainability are positioning themselves for durable growth.
“What’s Resonating with Firms Right Now: From AI Curiosity to Strategic Clarity”
From Curiosity to Clarity: What Firms Are Actually Asking For
Across the profession, firm leaders are expressing a consistent frustration with AI conversations that generate more noise than progress. What resonates most today is a structured, executive-level approach that starts with business problems—not tools—and results in a short list of high-impact, actionable use cases.
At LBMC, this demand led to the development of our AI Strategy Workshop, a working session designed to help leadership teams align AI initiatives with firm strategy, governance, and measurable outcomes. I lead these sessions alongside innovation strategist Charlie Apigian, with a focus on moving firms beyond AI discussion and experimentation into practical decision-making and execution. Rather than focusing on theoretical possibilities, the discussions center on where AI can realistically improve capacity, quality and economics within the firm today and what should intentionally wait.
For readers interested in this approach, Charlie Apigian’s book, AI Reimagined, offers additional perspective on reframing AI as a business transformation discipline rather than a technology trend.
The People Question: Where AI Is Reducing Work—and Where Humans Still Matter Most
What accounting work will see substantially less human involvement by 2026?
The most immediate impact of AI through 2026 will be felt in back-office processes that consume significant manual hours but deliver limited strategic value. In accounting, this includes accounts payable processing, routine journal entries requiring categorization, extraction of data from bank and financial statements and large portions of tax research.
These are areas well suited for automation with a human-in-the-loop model where AI performs the heavy lifting, and professionals focus on validation, exception handling, and judgment. The result is not the elimination of roles, but the elimination of low-leverage work that historically constrained capacity.
What will still be done primarily by humans?
High-trust, complex tax and audit work will remain firmly in human hands. When stakes are high—regulatory exposure, reputational risk, or material financial outcomes—clients want accountability, experience and judgment.
AI excels at the routine, standard and repeatable. It does not yet replace the nuanced reasoning required to interpret ambiguity, apply context, or advise clients through complexity. In accounting, trust is the product—and trust is still human.
Is there an AI compensation premium?
At this point, AI fluency is becoming table stakes rather than a differentiator. This is as much a leadership issue as a compensation one. Leaders are responsible for equipping their teams with the tools required to execute the firm’s vision; professionals are responsible for using those tools effectively. A well-qualified AI user should reasonably be 10–20% more efficient than a non-user. From a pure value-generation standpoint, that productivity delta provides a rational benchmark for compensation differentiation—not because of AI itself, but because of the outcomes it enables.
How AI has changed my own work
AI has delivered meaningful leverage in areas such as contractual language review, engagement negotiation support and communication quality. Nearly every email I send is now proofread and refined through AI. More importantly, AI has accelerated my ability to research and evaluate AI-enabled efficiencies as I work through complex client challenges.
The time saved has not reduced the scope of my role but rather it has expanded it. That time is reinvested in strategy, problem-solving, and client value creation.
Firm Dynamics: Cost, Competition and Compounding Advantage
AI is already reshaping competition across the profession. The Big 4 are investing heavily, while software providers and startups continue to automate core accounting processes. Early research suggests AI delivers a compounding effect: firms that build successive use cases generate exponentially greater ROI than those that experiment in isolation.
Firms that allocate budget, establish governance and move forward with a clear AI strategy will be better positioned to manage disruption and capitalize on it.
From a cost perspective, automation-driven AI adoption should reduce time and enable more disciplined pricing. However, price pressure on routine services is inevitable as automation becomes widespread. What AI will not replace is the human judgment required to interpret complexity, apply context and assume accountability.
This dynamic may also accelerate market consolidation, with larger firms pushing downstream to capture share as efficiency gains reshape pricing structures across the market.
Will AI Replace Accountants or Skills?
The more accurate framing is this: AI replaces skills, not people. If a professional spends most of their day on routine data entry, and that work is automated, the skill is displaced, not the individual. Displacement only occurs when a role fails to evolve.
Entry-level accountants will increasingly spend less time on rote tasks and more time apprenticing in complexity—learning judgment, client communication, and industry context. Firms that intentionally redesign early-career experiences will develop stronger professionals faster.
What would cause true concern?
True concern would require the arrival of artificial general intelligence—systems capable of handling ambiguous, end-to-end problem-solving with accountability and judgment. Today, AI still struggles with real-world execution. If it cannot reliably operate a vending machine, it cannot yet replace professional accountability.
The real risk is not AI itself but standing still. Competitors who use AI to become more productive, more differentiated and more cost-effective will apply pressure through pricing and value. That competitive reality— not technological fear—is what firm leaders must confront.
From Insight to Revenue—and Referrals
AI is not a technology initiative. It is a leadership imperative. Firms that start with the problem, apply AI with discipline, and align adoption to revenue, client experience and talent development will earn trust in a market saturated with hype.
That trust drives deeper relationships, stronger referrals and durable growth. The firms that win in 2026 will not be those asking whether AI belongs in accounting—but those who have already decided how to lead with it.
About the Author
Based in Nashville, Jon Hilton is the Shareholder-in-Charge of LBMC’s AI Practice and a seasoned Data Scientist with deep experience in artificial intelligence, machine learning, and data strategy.
