SMB Finance Is Exploring AI - But Few Are Actually Using It
Three-quarters of CFOs say they're actively exploring AI tools. Only one in ten is using them today. That gap tells the story: interest is high, execution is lagging.
A new poll from Accounting Seed - spanning 128 CFOs, controllers, finance staff, and IT/operations admins - found 63% are exploring AI but only 16% are actually using it. About 60% of respondents work at companies with 51 to 500 employees.
What the Survey Shows
Among CFOs, 10% report current AI usage while 77% are exploring options. Just 12% of organizations have reached "advanced adoption," using both automation tools and AI assistants.
More than a third (34%) still rely mostly on manual processes with no automation or AI in place. And 14% have no plans to use AI in accounting.
Why Adoption Is Stuck
"Many finance teams want to adopt AI but don't know which solutions are credible, haven't built the foundational data that AI requires, and lack the internal alignment to move forward," the report notes.
AI adoption appears strongest among IT and Salesforce administrators. That creates a real risk: technology-first projects that don't tie back to finance outcomes. Among those with no plans to adopt AI, the top barrier is leadership buy-in - finance needs a bigger voice early, and technical teams need clearer workflows to solve.
What Finance Teams Want AI To Solve
Asked for the top problem AI should address, 56% chose reducing time spent on repetitive tasks - three times more than any other option. Most respondents (84%) spend at least a quarter of their time on manual work; a third (33%) spend more than half their time on it.
Key nuance: many routine finance tasks are better handled by automation than AI. Payment processing, bank reconciliation, and approval routing often need well-designed workflows with clear rules - not machine learning.
A Pragmatic Plan for SMB Finance Leaders
- Start with a process audit: List high-volume, rules-based tasks and estimate hours spent, error rates, and cycle time. Prioritize the biggest manual time sinks.
- Tighten your data foundation: Standardize the chart of accounts, vendors, and coding rules. Clean data beats clever models.
- Automate first, then add AI: Use workflow/RPA inside your finance stack for AP, bank recs, and approvals. Layer AI where judgment, pattern detection, or language helps (invoice data capture, anomaly detection, forecast commentary, policy Q&A).
- Align finance and IT: Define outcomes in finance terms (days to close, manual hours removed, exception rate). Assign owners and timelines.
- Run 90-day pilots with clear baselines: Track before/after on close time, errors, exceptions, and hours saved. Kill or scale based on proof.
- Address risk and controls: Lock down access, protect PII, and document model usage and human review. See the NIST AI Risk Management Framework for guidance (NIST AI RMF).
- Upskill the team: Train accountants on automation and prompt skills; equip leaders to drive governance and buy-in.
Next Steps That Move the Needle
If your team spends 25% to 50% of its time on manual work, you don't need to start with AI. Start with clean data and solid workflows. Then add AI where it meaningfully reduces manual review or accelerates analysis.
For hands-on guidance, see the AI Learning Path for Accountants and the AI Learning Path for CFOs. Build the foundation, prove the value, and keep the scope tight. That's how exploration turns into results.
Your membership also unlocks:
