Most CPA firms are not struggling to access AI. They are struggling to use it in a way that actually changes how work gets done.
Over the past year, firms have invested in tools, tested use cases and encouraged teams to experiment. On paper, adoption looks like it is happening. In practice, many firms are stuck in the same place. Usage is inconsistent. Workflows remain largely unchanged. Leadership is left wondering why the investment has not translated into meaningful impact.
It is easy to assume the issue is budget, training or hesitation. In most cases, it is something more fundamental.
Most firms don’t lack access to AI tools. They lack the data that those tools can actually use. Recent research from AICPA and CIMA points to data quality and availability as some of the most common barriers to AI adoption in accounting firms.
That gap shows up quickly in day-to-day work.
A staff member tries to use AI to summarize client documentation, but the files are spread across email, shared drives and multiple systems. A partner experiments with drafting a response using Copilot, but the underlying data is incomplete or inconsistent. A firm rolls out AI licenses, only to find that usage drops off after the first few attempts.
The issue is not the tool. It is the environment the tool depends on.
Firms don’t have an AI problem. They have a data problem.
The Myth vs. Reality of AI Adoption
There is a growing assumption that firms are behind on AI because they are moving too slowly. That is not what we are seeing. Many firms have already taken the first steps. Licenses have been purchased. Tools are available. Some level of experimentation is happening across teams.
What isn’t happening is follow-through.
Copilot sits unused after the initial rollout. ChatGPT is used individually but never becomes part of firmwide workflows. Early enthusiasm fades when results are inconsistent or underwhelming.
At the same time, the technology itself is moving quickly. AI capabilities are now being embedded directly into tax, audit and advisory platforms, making access easier than ever.
The bottleneck is not availability. It is the firm’s ability to support AI with usable, connected and trusted data.
Many firms are finding that simply providing access to AI tools is not enough to drive meaningful change. As we’ve explored in a previous post, adoption requires more than access. It requires structure, guidance and alignment with real workflows.
The Data Problems Getting in the Way
If AI adoption is stalling, the root cause is rarely a lack of interest or investment. It is usually the condition of the data behind the scenes. This is not unique to accounting. Research from Deloitte shows that data challenges, not budget or tooling, are the primary barrier to scaling AI across organizations.
CPA firms feel this more acutely because of how their work is structured. Client data moves across multiple systems, formats and workflows, often without consistency or clear ownership.
Four issues tend to surface repeatedly:
1. Data is fragmented across systems
Client information rarely lives in one place. Tax software, audit platforms, document management systems, email and time and billing tools all hold pieces of the same story.
For AI, that fragmentation is a problem. AI cannot connect insights across systems that are not connected in the first place. What looks like a single client relationship to a partner often exists as disconnected data points behind the scenes.
2. Data is inconsistent and unstructured
Even within a single system, data is rarely standardized.
File naming conventions vary by team or individual. Key information is buried in PDFs, email threads or spreadsheets. Similar documents are stored in different formats with no consistent structure.
AI depends on patterns. When data is inconsistent, outputs become unreliable. That is often the moment where confidence in AI starts to break down.
3. Data access is unclear or restricted
In many firms, access is either too broad or too limited, with little consistency in between.
Permissions evolve over time. Roles change. Temporary access becomes permanent. At the same time, security concerns make firms hesitant to allow AI tools to interact with sensitive data at all.
The result is a standstill. Either AI cannot access what it needs, or firms are not comfortable letting it try.
4. Data governance is informal or nonexistent
Behind all of this is a lack of clear ownership. Firms are left asking: “Who is responsible for data quality?” “Who defines structure?” “Who decides how long information is retained or how it is categorized?”
In many firms, those answers are unclear. Data management happens passively, not intentionally.
AI does not fix that. It exposes it.
Why This Matters More Than Firms Think
When AI initiatives stall, it is easy to blame the tool. In reality, the problem usually starts much earlier.
Poor data leads to inconsistent outputs. Inconsistent outputs lead to low trust. And once trust is lost, adoption drops off quickly. That is how firms end up saying, “We tried AI, and it didn’t work.”
Gartner has consistently pointed to poor data quality as a primary reason AI initiatives fail to deliver expected outcomes. For CPA firms, the impact is not just technical. It is operational.
Staff stop using tools that feel unreliable. Partners lose confidence in the results. Investments fail to translate into capacity gains or efficiency improvements. This pattern shows up across multiple areas of firm operations. When foundational issues are left unresolved, they tend to surface under pressure, especially during peak periods like busy season.
AI does not fail on its own. It reflects the quality of the inputs behind it.
What “AI-Ready Data” Actually Looks Like
AI-ready data does not mean perfect data. It means data that is usable, accessible and consistent enough to support real work.
In practice, that looks like:
- Systems that are connected or intentionally integrated
- Clear and consistent naming conventions across files and documents
- Defined access controls aligned to roles
- Shared understanding of where key data lives
- Ownership of data quality and structure
Firms do not need to solve everything at once. But they do need to move from passive data management to intentional data discipline.
For firms thinking about how this plays out during peak workloads, it is worth considering how data access, structure and control impact day-to-day execution under pressure.
Where Firms Should Start
The fastest way to stall progress is to treat this as a massive transformation effort.
It is not.
Firms that make progress with AI usually start small and stay focused.
- Audit where your data lives today
- Identify one workflow to improve, such as client onboarding or tax document collection
- Standardize how data is stored and labeled in that area
- Clarify who has access and why
- Build from there
This approach creates early wins without overwhelming teams or disrupting existing workflows.
How IT Leadership Should Be Thinking About This
AI readiness is not just a technology issue. It is an operational one. The conversation needs to shift from “What tools should we buy?” to “Is our data in a condition where those tools can actually help?”
For IT leaders, this means focusing less on adding new platforms and more on strengthening the foundation those platforms depend on. For firm leadership, it means recognizing that data discipline is not back-office work. It directly impacts efficiency, capacity and client service.
The firms that move forward with AI are not the ones with the most tools. They are the ones with the clearest, most usable data. This is where many firms begin to rethink how their technology environment is structured, not just to support AI, but to reduce risk, improve visibility and create more predictable operations.
The Bottom Line
AI has the potential to meaningfully change how CPA firms operate. But that change is not complete just because you’ve purchased the tool. Firms that invest in AI without addressing their data often find themselves stuck in pilot mode, testing capabilities without seeing real impact.
Firms that focus on data first move faster, with more confidence and better results.
AI success doesn’t start with the tool. It starts with the data behind it.
Netgain works with CPA firms to help clean up the foundation AI depends on, from data structure and access to overall IT strategy. If you want a second set of eyes on where to start, you can learn more or start a conversation here.