A useful reality check for small-business AI
Artificial intelligence is no longer an experiment reserved for large enterprises. A July 15 research brief from the Federal Reserve Bank of San Francisco reports that nearly 40% of respondents to the 2024 Small Business Credit Survey were either using AI or planning to use it. The finding is important not because every company should rush to buy another tool, but because AI is becoming part of ordinary operating decisions.
The brief also shows that adoption is uneven. Businesses described uses ranging from basic writing assistance to integrated systems. Common applications included productivity work, marketing and search optimization, customer service, analytics and forecasting, software development, visual design, and custom tools. That breadth confirms a point we see in consulting work: “AI adoption” is not one project. It is a portfolio of workflow changes with different risks and returns.
What happened
The San Francisco Fed analyzed qualitative responses from small businesses in the 2024 survey. Enthusiasm existed alongside opposition and uncertainty. Interested firms cited policy or regulatory constraints, tool costs, limited time for training and upgrades, and gaps in implementation knowledge. Other firms intentionally stayed away because of accuracy, intellectual-property, human-interaction, or relevance concerns.
This is a better decision framework than treating adoption as a race. A company using AI in five disconnected apps without controls may be less mature than one using it in one carefully measured process. The relevant questions are whether the workflow is repeatable, whether source data is reliable, whether a person remains accountable, and whether the result improves a business metric.
Why it matters for growing businesses
SMBs have less room for failed technology projects, but they can often redesign processes faster than large organizations. That makes focus a competitive advantage. Start with work that is frequent, rules-based, and easy to review: classifying incoming requests, drafting standard communications, extracting fields from documents, preparing management summaries, or suggesting next actions in a CRM.
The barriers identified by the survey are connected. Weak data increases accuracy concerns. Unclear policy encourages shadow AI. A lack of training causes employees to paste sensitive information into consumer tools. Disconnected systems create manual copying that erodes the time savings. Solving these issues together turns an isolated tool into an operating capability.
Practical action checklist
- Inventory current AI use, including free tools employees adopted independently.
- Pick one workflow with a clear owner, baseline time, error rate, and desired outcome.
- Classify the data involved and prohibit sensitive inputs in unapproved services.
- Define where human review is mandatory and log important decisions.
- Connect approved tools to the ERP, CRM, or document system instead of creating another silo.
- Run a four-to-six-week pilot and compare actual quality, cycle time, and cost with the baseline.
- Train the team on both effective prompting and verification—not prompting alone.
A simple portfolio view
Leadership can group candidate uses into three lanes. “Assist” cases draft or summarize while a person remains fully in control. “Automate” cases execute a defined step under deterministic rules. “Advise” cases produce recommendations that may influence a decision. Each lane needs a different approval threshold and review cadence. Keep a one-page register showing the workflow owner, data used, vendor, expected benefit, failure mode, and fallback. Review it quarterly and close pilots that do not produce measurable improvement. This prevents a collection of subscriptions from being mistaken for a strategy and gives employees a clear route for proposing useful experiments.
How VDS Consultoria can help
VDS Consultoria helps growing organizations move from scattered experimentation to governed automation. We map workflows, identify high-value use cases, connect AI services to Odoo and other business systems, define practical controls, and build dashboards that show whether the change is working. The goal is not maximum AI usage. It is better service, fewer repetitive tasks, and decisions that remain explainable and accountable.
Source
Federal Reserve Bank of San Francisco, “AI Adoption Among Small Businesses: Qualitative Insights from the Small Business Credit Survey,” July 15, 2026: https://www.frbsf.org/research-and-insights/publications/community-development-research-briefs/2026/07/ai-adoption-in-small-businesses-2024-sbcs/